Foundations and Trends
R

in
Technology, Information and Operations Management
Vol. 3, No. 1 (2009) 1–100


c 2010 A. Apte
DOI: 10.1561/0200000014

Humanitarian Logistics: A New Field of

Research and Action
By Aruna Apte

Contents

1 Introduction 3

2 Humanitarian Logistics 6
2.1 The Need 7
2.2 Lessons Learned 10
2.3 Disaster Classification 13
2.4 Disaster Response and Humanitarian Relief 16

3 Design of Humanitarian Logistics 18

3.1 Military, Commercial and Humanitarian Logistics:
A Comparison 19
3.2 Supply Chain Considerations 21
3.3 Humanitarian Logistics as a Supply Chain 27

4 Supply Chain Issues in Humanitarian Logistics 29

4.1 Preparation 31
4.2 Disaster Response 45
4.3 Relief Operations 58
5 Organizational Issues in Humanitarian Logistics 66

5.1 Collaboration 66
5.2 Information and Knowledge Management 68
5.3 Training and Education 69
5.4 Role of Donors and Donations 69
5.5 Risk Management 70

6 Conclusions and Potential for Future Research 72

6.1 Examples of Actual Practice 72

6.2 Challenges Today in Humanitarian Logistics 77
6.3 Future Research 83

Acknowledgements 88

References 89
Foundations and Trends
R
in
Technology, Information and Operations Management
Vol. 3, No. 1 (2009) 1–100


c 2010 A. Apte
DOI: 10.1561/0200000014

Humanitarian Logistics: A New Field of

Research and Action

Aruna Apte

Graduate School of Business and Public Policy, Naval Postgraduate School,

555 Dyer Road, Monterey CA, 93943, USA, [email protected]

Abstract
Recent natural disasters such as the earthquake in Haiti, Hurricane
Katrina in the United States, tsunami in the Indian Ocean, the earth-
quake in Pakistan, and numerous humanitarian challenges arising from
such conflicts as that in Sudan have exposed the shortcomings in plan-
ning for disasters. In addition to the natural disasters, the homeland
security issues related to domestic as well as international terrorism
have increased the fear factor and have made ‘readiness’ the principal
priority. Humanitarian logistics is a critical element of an effective disas-
ter relief process. The objective of this monograph is to discuss research
issues and potential actions surrounding the new field of humanitar-
ian logistics. We define humanitarian logistics as that special branch
of logistics which manages response supply chain of critical supplies
and services with challenges such as demand surges, uncertain sup-
plies, critical time windows in face of infrastructure vulnerabilities and
vast scope and size of the operations. We survey case studies to learn
from the past experience and review analytical models from the liter-
ature to understand the state-of-the-art in humanitarian logistics. We
recommend further research in the fields of operations management
and operations research to improve the effectiveness and efficiency of
humanitarian logistics. We conclude that though humanitarian logis-
tics is inherently chaotic and complex, and it is difficult to do research
in this area, the complexity and obstacles can be dealt with by the
researchers.

Keywords: Humanitarian; logistics; disaster; emergency; relief;

response; supply chain; analytical models; preposition-
ing; facility location; distribution; evacuation; inventory;
decision making.
 1
Introduction

A significant proportion of the world’s population has suffered in recent

years as a result of disasters — both natural and manmade. Human-
itarian responses to the 2004 tsunami in the Indian Ocean, the 2005
earthquake in Pakistan, various hurricanes in the United States, the
conflict in Sudan, and the spread of HIV/AIDS in Africa have largely
been neither effective nor efficient. A recent example of this is the 2010
earthquake in Haiti and the devastation after that. The reasons are
many, but are partly attributable to the sheer size and scope of such
disasters. As reported by the Center for Research on the Epidemiology
of the Disaster (CRED), between 1974 and 2003 there were 6,637 nat-
ural disasters worldwide that affected 5.1 billion people and accrued
$1.38 trillion (US$) in damage [56]. In 2007 alone, 106 natural disas-
ters killed 17,000 people and affected 201 million more [107]. Between
May and September 2007, the Red Cross responded to natural disas-
ters in 18 countries in Africa, 16 in the Americas, 13 in Asia, and 10 in
European Nations [69].
As a result of the Indian Ocean tsunami in 2004, the overall bud-
gets for humanitarian efforts have increased — and this in turn has
brought increased attention to humanitarian logistics. It is commonly

3
4 Introduction

agreed that disaster relief has a large logistics component [82, 162, 164].
With the 2004 budgets of the top 10 humanitarian agencies exceeding
$14 billion in total, the logistics of aid has attracted recent scrutiny
[155]. Consequently, humanitarian logistics has become a topic of inter-
est to both academics and practitioners [82]. An effective and efficient
humanitarian response depends “on the ability of logisticians to pro-
cure, transport and receive supplies at the site of a humanitarian relief
effort” [153].
The objective of this monograph is to take a comprehensive look
at the issues surrounding humanitarian logistics through the life cycle
of the disaster spanning three stages of operations: preparation, dis-
aster response, and humanitarian relief. We define humanitarian logis-
tics from different perspectives, survey case studies to delineate lessons
learned, and review analytical models from the literature to understand
the state-of-the-art in humanitarian logistics. We then recommend fur-
ther research in the fields of operations management and operations
research to improve the effectiveness and efficiency of humanitarian
logistics.
This monograph consists of six sections, including the current intro-
ductory section. In the next section, we review various aspects of
humanitarian logistics to provide a comprehensive definition. We dis-
cuss the need for humanitarian logistics, review the lessons learned in
past humanitarian logistics efforts, develop a classification of disasters,
and distinguish between disaster response and humanitarian relief. In
Section 3 we discuss the factors important for the design of humani-
tarian logistics using traditional concepts of logistics and supply chain
management. Specifically, we compare humanitarian logistics with the
military logistics as well as commercial supply chains.
Sections 4 and 5 describe supply chain issues and organizational
issues in humanitarian logistics. In Section 4 we review analytical mod-
els related to humanitarian logistics. In reviewing analytical models we
focus on efficiency and effectiveness of various aspects of humanitar-
ian logistics. This review includes models that have been published in
extant research literature as well as those that have been proposed
in various working papers known to the author. We should point out
that although this monograph is intended for all audiences, academics
 5

as well as practitioners, the academic community may find the dis-

cussion of analytical models in Section 4 particularly useful in their
research efforts. Section 5 discusses organizational issues, such as col-
laboration among players, flow of information, knowledge management,
etc., that are important factors in humanitarian logistics. Motivation
for discussing the literature in these two sections is to present an infor-
mal survey of the current research.
In Section 6, the last section, we present case studies of current
practices to place this academic research in the context of reality. We
believe this will provide the broader picture of humanitarian logistics.
In the same section we also describe the challenges as seen by the
officials in the field. The case studies combined with challenges highlight
the research needed in humanitarian logistics, a new field of research
and action.
 2
Humanitarian Logistics

Response to a disaster must be tailored to the characteristics of the dis-

aster. A disaster is defined by Federal Emergency Management Agency
(FEMA) as an event that causes 100 deaths or 100 human injuries or
damage worth US$ 1 million. Preparing for a humanitarian response
predominantly involves managing the logistics of potential responses
[156, 164]. In order to better understand this process, we must define
humanitarian logistics. To this end, in this section we first explain the
need for logistics. Furthermore, in order to understand humanitarian
logistics we also need to understand disasters. Hence we present a clas-
sification of disasters based on time and location. We then review the
lessons learned so far to gauge the scope of humanitarian logistics.
Finally, we distinguish between disaster response and humanitarian
relief to establish the definition of humanitarian logistics.
The term, “Logistics,” has different meanings to different organi-
zations and people. The goal of operational logistics in the military,
for example, is to sustain military operations and bridging the gap
between strategic logistics (e.g., national stockpiles, industrial infra-
structure, and power-projection capabilities) and tactical logistics (e.g.,
unit-level sustainment) [83]. In the business sector, logistics is defined
as a “planning framework for the management of material, service,

6
 2.1 The Need 7

information, and capital flows and includes the increasingly complex

information, material, communication, and control systems required in
today’s business environment” [164]. In contrast, most humanitarian
organizations such as the World Food Program (WFP) and Medicines
Sans Frontiers (MSF) agree that humanitarian logistics is “the process
of planning, implementing, and controlling the efficient, cost-effective,
flow and storage of goods and materials, as well as related information,
from point of origin to point of consumption for the purpose of meeting
the end beneficiary’s requirements” [154]. In maintaining and increas-
ing efficiency and effectiveness of this logistics, one important fact to
remember is that the three principles of “humanitarian space”, human-
ity, neutrality, and impartiality, have to be present during the strategic,
tactical, and operational stages of humanitarian operations [158].

2.1 The Need

History and research suggest that humanitarian logistics should be
more evidence-based. One of the ways to verify is if resources allo-
cated for planning or responding to any disaster are justified by data.
However, “Data on disaster occurrence, their effect upon people and
cost to countries remain at best patchy” [56]. The key sources of such
data in recent years have been insurance companies as well as special-
ized agencies such as the World Food Program (WFP) and the World
Health Organization (WHO). The recent increase in the occurrence of
disasters shown in Figure 2.1 can possibly be explained by the increase
over time in the reporting of small disasters. Other plausible reason is
the population increase fueled by migration to disaster prone areas.
Similar to under-reporting of disaster occurrences, the details
regarding economic losses are also frequently under-reported. Difficul-
ties in arriving at accurate estimates of economic loss and the absence of
internationally accepted methods of damage assessment [56] are partly
responsible for the lack of data on economic losses caused by disas-
ters. In any event, the devastating effect of disasters on economies
and human health has prompted investigations into the underlying
causes and their remedy resulting in the need for humanitarian logis-
tics. For example, one such remedy in reducing human suffering in
8 Humanitarian Logistics

Fig. 2.1 Natural disasters 1974–2003 (Source: Guha-Sapir et al., [56]).

natural disasters is to proactively prepare for and pre-position supplies

in best possible locations. Economy of the affected area depends on
transportation, so one solution may be securing infrastructure issues
through long-term actions. These strategic issues in logistics are major
challenges in face of funding problems.
But humanitarian agencies face a number of operational challenges
as well. For example, though their primary mission is to respond to
disasters as they occur, humanitarian agencies must also satisfy their
donors, without whose substantial pledges the affected communities
could not be served [115, 164]. Additionally, many of the distributed
supplies are donations received to the humanitarian agencies instead
of being purchased directly by them. As a result, humanitarian agen-
cies have the additional task of managing donated supplies which may
or may not be necessary or suitable for the given disaster. The prob-
lems experienced with the suitability of donated supplies in case of the
Pakistan earthquake in 2005 are presented in Exhibit 1. Humanitar-
ian agencies must streamline the in-flow of donations so that “differ-
ing or overlapping aid” do not result in “oversupply or inappropriate
supply” [107]. The need to suitably handle genetically modified food,
about-to-expire medications, and substantial amounts of unused sup-
plies create further challenges.
 2.1 The Need 9

Exhibit 1: Pakistan Earthquake (2005)

Greg Mortensen, the best-selling author of Three Cups of Tea, has
worked extensively in Pakistan and Afghanistan for humanitarian
purposes such as building schools. After the Pakistan earthquake he
noticed women burning fancy tweed jackets as fuel for cooking food.
It seemed there was enough warm clothing to go around but no fuel
for cooking.
The tents donated by kind individuals for shelter came with no
instructions. And in some instances, even if the instructions were
there, most of the victims could not understand the language. Folks
did not know how to put those tents up. Moreover, some key pieces
were missing. After resolving all the issues when the tents were up,
large families gathered in one tent around the fire to stay warm. The
tents caught fire and instead of being any help they became a deadly
hazard.
Source: Mortensen [111].

According to the Munich Reinvestment group, the real annual eco-

nomic losses have grown from US$75.5 billion in 1960s to US$659.9
billion in the 1990s. The forecasts are for five fold increases in the
number of natural and manmade disasters in the next 50 years [155].
In light of this, the field of humanitarian logistics is gaining more inter-
est and has seen a significant increase in research in the recent years.
Although response supply chains for humanitarian causes are arguably
among the “most dynamic and complex supply chains in the world”
[154], proper logistics preparation before a disaster strikes can be done
by coordinating processes, technologies, and communications capabil-
ities. Such actions can improve the effectiveness and efficiency of the
response supply chains, and thus that of the humanitarian agency’s
response. Academic research, based on input from practitioners and
using operations management and operations research analysis, is crit-
ical in bridging the gap between logistical expertise and humanitarian
relief. Research in areas of humanitarian relief and disaster response
belongs to a special subset of the broad area of logistics.
An onslaught of recent disasters, both natural and manmade,
and the fear of those that can potentially occur, have led to several
10 Humanitarian Logistics

research projects. The resulting articles typically discuss lessons

learned, describe analytical models for humanitarian supply chain, and
explore conceptual boundaries between various relief operation topics.
Taken as a whole, they provide a substantial foundation for improving
humanitarian logistics. For example, improved decision-making models
can help humanitarian organizations and first responders refine their
strategy for planning, manage their operations for prepositioning, stor-
age and distribution, and sustain aid both during and after a crisis.

2.2 Lessons Learned

All recent disasters have had both short-term and long-term conse-
quences, and they provide us with a wealth of lessons for the future
operational planning and execution of humanitarian logistics. One of
the most important lessons learned is that money is not a panacea,
as evidenced in the aftermath of the tsunami in Ache, Indonesia (see
Exhibit 2). The enormous quantities of food, medicine, and water that
were made available as aid from outside the region (response to the
disaster) could not reach the disaster victims due to non-existent dis-
tribution channels and prepositioned assets (lack of preparation) with
difficulties in the last mile distribution (relief operations).

Exhibit 2: Tsunami Devastation in Banda Aceh 2004

Banda Aceh is the regional capital of Aceh, Indonesia. After the
tsunami in the Indian Ocean there were enough supplies donated by
the world’s richest countries but there was only one airstrip and one
forklift in Banda Aceh. Having available huge amounts of supplies
was of less consequence since there were no means for distributing
them. Money available from donations could neither be employed at
the eleventh hour to build an airstrip nor to buy forklifts at the last
minute. So the officials there had difficulty delivering food, water and
medicines to the people who survived the disaster but needed critical
commodities to carry on. In short, “the broader point is that money
is not a panacea” [62].
Source: Heidtke [62], The Economist Global Agenda [152].
 2.2 Lessons Learned 11

The horrific images of the devastation from Hurricane Katrina and

the reports of how public officials failed to mitigate the resulting dam-
age are plentiful. However, one of the lessons learned was that govern-
ment officials were not the only ones ill-prepared to face that disaster.
Supply chains of many manufacturers and shippers were interrupted as
well. Research done by logistics managers immediately following Kat-
rina indicates that the majority of companies had been holding far more
inventory than in the previous era of “lesser inventory the better” [27].
However, despite higher inventory levels, the supply chains of these
firms were not positioned to weather a disaster of such magnitude.
Hurricane Katrina highlighted the fact that logistics managers must
consider a multitude of factors relevant to their supply chain, includ-
ing developing a contingency plan for warehouse locations, forecasting
demand to estimate the amount of inventory to be held, determining
the appropriate amount of flexibility in their distribution channels, etc.
Whether to hold more inventories to reduce lead time or not protect
against destruction is something that can be learned from the past.
However, such decisions are further complicated by the fact that the
more global the supply chain, the more difficult are the contingency
planning, humanitarian efforts, transportation, and distribution.
In a response supply chain, delivery during the last mile is always an
issue. When Telegraph was developed over a century ago, it revolution-
ized the way people communicated. Although it was a major techno-
logical breakthrough, the system still required a runner to deliver the
actual message from the local telegraph office to the ultimate recip-
ient — a requirement known as the ‘last mile problem.’ Today, the
humanitarian agencies are facing the same last mile distribution prob-
lem: how to deliver the right supplies to the right population at the right
time? Frequently, the supplies are plentiful but they cannot be effec-
tively deployed because of logistical problems in the response supply
chain. For this last phase of humanitarian aid, scheduling of deliver-
ies, routing of delivery vehicles, effectively managed supply inventory,
and efficient transportation of these supplies play a crucial roles in this
response.
Studies in logistics conclude that infrastructure such as distribu-
tion centers, warehouses, and medical clinics should be established in
12 Humanitarian Logistics

locations based on proximity to high-population areas and transporta-

tion hubs. Facility location is of utmost importance for distribution of
critical supplies and services under uncertainty.
In addition to location of distribution centers, for the last mile distri-
bution, warehousing strategies involving prepositioning of needed sup-
plies is also important in many other complex humanitarian logistics
systems. One such lesson learned was through the response to contin-
ued humanitarian relief for the millions of displaced people in South
Sudan. The United Nations Children’s Fund (UNICEF) established
Operation Lifeline Sudan (OLS) to operate and provide airlifts of food
and critical supplies from Lochichoggio, Kenya. Because of presence of
OLS and the security it offers, many non-governmental organizations
(NGOs) pre-positioned relief items in the warehouses there. Here the
inventory management played an important role in warehousing [15] in
terms of order quantities, reorder points, and types of demand.
However, even with sound prepositioning, warehousing, inventory
management, transportation, and last mile distribution, the unex-
pected surges and the diversity of demand require the processes to
be agile to ensure that the aid can reduce the victim suffering. “Rapid
deployment on demand” [164] is an essential strategy that can lead to
successful humanitarian relief effort.
In addition to the deployment of supplies and services, another les-
son learned due to the evacuation fiasco after Hurricane Katrina in
the United States is rapid mass evacuation planning. Unfortunately,
during Katrina, “beginning with the evacuation orders before the hur-
ricane landfall, some public officials did not know what those right steps
might be” [114]. Some of the lessons learned here are that there exist
specific vulnerabilities of evacuees such as race and culture [38]. It also
suggests the need for evacuation planning [8, 114, 138, 167].
Mass evacuation is not possible if there is no coordination, co-
operation, and collaboration amongst parties involved. This can be
accomplished through the network of community-based volunteers and
staff members. It is essential not just for evacuation but for support-
ing all stages of the humanitarian logistics, planning, deployment, and
recovery effort. One of the major lessons learned from recent disasters is
 2.3 Disaster Classification 13

that without the collaboration between partners — internal and exter-

nal to the system, military and civilian, private sector and non-profit
organizations — execution of disaster response is likely to be far from
adequate. In fact, without such collaboration humanitarian operations
may be derailed by a “multiplicity of agencies and governments” [135].
Such efforts from staff members and especially volunteers of human-
itarian agencies cannot be efficient if the personnel involved are not
trained properly. The need for training and education of first respon-
ders has been demonstrated in various crises. In addition, standards
must be developed “to ensure that emergency planning, field opera-
tions and personnel training are conducted rigorously” [2]. In order
to appreciate these outcomes and act on them in the future we need
to understand this process of humanitarian logistics for which we first
need to understand disasters.

2.3 Disaster Classification

Disasters can be classified based on the specific characteristics of dis-
asters including the speed of onset, slow or sudden; and the source of
disaster, natural or manmade [43, 164]. It is customary to classify dis-
asters as natural versus manmade. Especially since between 1982 and
1994, 97% of humanitarian operations were devoted to disaster relief
for manmade disasters [164]. However, the funding typically flows to
the disaster that gets media exposure or strikes a nerve. The level of
resources available in responding to such manmade disasters can be sig-
nificantly different. Malice and intelligence in planning make manmade
disasters difficult to anticipate and hence problematic for humanitar-
ians agencies in several respects, such as safety and security, in addi-
tion to basic logistics operations. Yet, a classification based on man-
made versus natural disasters does not achieve substantial inference
for research in terms of operations. In either case it is a disaster which
needs to be responded to. Therefore we do not focus on this type of
classification.
Using the level of difficulty potentially present in humanitarian
operations as the criteria we classify disasters based on time and loca-
tion. This is illustrated in Figure 2.2. Based on this classification we
14 Humanitarian Logistics

can determine the factors that play important roles in disaster response
or humanitarian relief. This is not to say that providing aid in case of
localized slow-onset disaster is always easy but level of difficulty is lower
since it is possible to prepare for such disasters. Therefore, it helps if we
also pay attention to the distinction between three stages of humani-
tarian logistics operations: preparation, disaster response, and ongoing
humanitarian relief. We believe that preparation and prepositioning in
anticipation of a disaster response, once the disaster strikes, and the
on-going relief operations play a key role in determining the level of
difficulty of humanitarian logistics operations.
Preparation can play an important role in the disasters classified
in the first quadrant in Figure 2.2. The disasters being localized with
possible lead time for preparation due to slow-onset provide this oppor-
tunity. In this case, the actual response is not likely to be that difficult
since there is time to catch up and the disaster itself is geographically
contained.
The second quadrant where the onset is slow but locations are dis-
persed, preparation can help but pre-positioning becomes challenging.

Fig. 2.2 Classification of disasters.

 2.3 Disaster Classification 15

The large and scattered geographical area takes substantial amount of

budget and coordination among players. Therefore, disaster response
may play an important role here. The relief process may also pose issues
due to multiple locations.
The sudden-onset disaster, even if localized, in the third quadrant
creates problems in all the three roles due to obvious uncertainties. Yet,
being localized, the level of operational difficulty is somewhat lower as
compared to that posed in dispersed and sudden-onset disasters from
the fourth quadrant.
In summary, the disasters that occur over a period of time or disas-
ters with slow-onset provide time for humanitarians to plan and prepare
for relief operations. On the other hand a disaster that strikes suddenly
can pose formidable problems for response since no organization can
prepare well for such event. Strategic planning such as asset preposi-
tioning and facility location may help to some extent but the disaster
response is the name of the game here. It also makes a difference in diffi-
culty of response or relief whether the disaster is localized or if it affects
large and numerous geographical areas. Effectiveness and efficiency of
transportation and distribution of critical supplies and services suffer
if the disasters are dispersed, such as the 2004 tsunami in the Indian
Ocean which affected many countries.
Thus, the localized, slow-onset, and natural disasters (seasonal or
non-seasonal) are at one end of the spectrum of the level of difficulty for
humanitarian logistics whereas dispersed, sudden-onset disasters are at
the other end of that spectrum. In between are the rest, for instance,
localized and sudden such as 2005 hurricane Katrina in United States or
the severe earthquake–magnitude 7.9 on the Richter scale — in Gujarat,
India in January 2001. Though localized and sudden, there is a marked
difference between these two disasters, one is seasonal and the other is
not.
An influential factor in natural disasters is whether the occurrence
is seasonal or non-seasonal. In the United States, tornadoes in so-called
“Tornado Alley” (states such as Texas, Oklahoma, and Kansas where
frequent tornadoes occur) and wildfires in California are annual events
encountered in summer months. Earthquakes, on the other hand, are
16 Humanitarian Logistics

non-seasonal. Evacuation plans can be in place for either of the situa-

tion but will differ based on this factor. These different types of seasonal
disasters, even if sometimes they are sudden, pose lesser problems than
those faced in non-seasonal disasters in terms of the three key stages of
humanitarian logistics. Classification of disasters and the three stages
in the life cycle of a disaster suggest that we examine the similarity
and distinction between disaster response and humanitarian relief.

2.4 Disaster Response and Humanitarian Relief

In a disaster response, there is immediate need for critical supplies but
usually there is limited information available about the requirements.
Hence, disaster response tends to become, to some extent, “reactionary
logistics” [143]. Time is a distinguishing factor in disaster response
since there is generally limited warning prior to the event and urgent
response is needed because suffering can increase rapidly in the absence
of timely aid. Humanitarian relief is an ongoing process of mitiga-
tion through disasters. It is important to note, however, that in the
case of seasonal disasters it is possible to plan for the preposition-
ing of assets and critical supplies, and to plan for the evacuation of
the affected population using information from previous disasters. Fur-
thermore, the population affected in such sudden-onset disasters —
natural or manmade — is normally concentrated in specific geographic
areas. There exist exceptions to this generalization, such as the 2004
tsunami in the Indian Ocean, which affected numerous countries, and
the 2008 terrorist attacks in Mumbai, India, which were widespread
and simultaneous although contained in a relatively small area of the
city. Disaster response also differs from humanitarian relief in that it
receives comparatively more media exposure as a result of its urgency
and sudden-onset.
Humanitarian relief, on the other hand, is an ongoing process for
slow-onset disasters with a long-term need for supplies. To a large
extent, relief requirements are known and relief organizations face rel-
atively long planning lead times. This is not to say that these oper-
ations are easy since there are many examples where providing relief
has posed numerous difficulties. Disasters such as famine in Sudan and
 2.4 Disaster Response and Humanitarian Relief 17

HIV/AIDS in Africa have affected populations in vast regions. In such

cases, sustained long-term relief operations are necessary. Thus, human-
itarian relief operations usually have sufficient time available in which
to plan their response before the relief process begins and therefore
humanitarian relief process lasts much longer than a disaster response.
A consequence of the long relief process is “exposure fatigue” with the
media tending to lose interest and thus limiting coverage, which ulti-
mately results in a diminished stream of donations.
Humanitarian logistics is an integral part of both disaster response
and humanitarian relief. We, in this monograph, primarily focus on
disaster relief logistics and the operations within this response sup-
ply chain. We define humanitarian logistics as that special branch of
logistics which manages response supply chain of critical supplies and
services with challenges such as demand surges, uncertain supplies,
critical time windows in face of infrastructure vulnerabilities and vast
scope and size of the operations. We regard humanitarian logistics as a
supply chain spanning the life cycle of a disaster. The operations in this
supply chain may be loosely divided into three stages: preparation, dis-
aster response, and ongoing humanitarian relief (sometimes known as
recovery). We use this classification for discussing supply chain issues in
humanitarian logistics throughout the monograph but more specifically
in the fourth section.
 3
Design of Humanitarian Logistics

In this section we discuss issues related to the design of humanitarian

logistics. We do this by first understanding the differences and simi-
larities between military, commercial, and humanitarian logistics. We
then frame the humanitarian logistics as the response supply chain
by describing various components and characteristics. We distinguish
logistics decisions at three levels of decision making: strategic, tactical,
and operational. Decisions made at the strategic level, such as research
and development of logistical capabilities, sourcing of supplies, deter-
mination of distribution policies, and build up of physical infrastruc-
ture, have long-lasting impacts [83]. Two important strategic issues for
humanitarian logistics are the building and maintenance of infrastruc-
ture and the prepositioning of assets and critical supplies. These issues
also encompass technological capabilities, inventory support for ware-
houses full of consumable and non-consumable supplies, building or
maintaining roads, bridges and airstrips. At the strategic level, human-
itarian logistics can be managed well if sufficient experience has been
gained from prior disasters and if sufficient funds are available. The
objective here is to develop capabilities and infrastructure to reduce
suffering while being cost-effective in disasters.

18
 3.1 Military, Commercial and Humanitarian Logistics: A Comparison 19

At the operational level, decisions pertain to the end-game being

played out in the affected areas. This stage of humanitarian logistics
potentially entails for example, evacuations of the affected population
or the last mile distribution of critical supplies and services, executed
with specificity and customization. An important operational detail
is applying such customization to supply inventory. The objective of
operational decision is to get relief to the affected population, quickly.
Tactical decisions in humanitarian logistics bridge the functional
gap between the strategic and operational levels. At this level, decisions
are made about gathering resources (inventory management) and then
transporting and deploying them as needed. The primary objective at
the tactical level is the real-time management of the supply chain, with
optimal inventory, routing, distribution, and scheduling of delivery of
the supplies, while keeping in mind that the supplies should be delivered
as soon as possible.
Analytical tools can be extremely useful in optimizing this response
supply chain at all levels. For example, models at strategic level for
resource allocation, and facility location; at tactical level for inventory
management, distribution and scheduling; and at operational level for
evacuation, lead time reduction and last mile distribution. This topic
is discussed at length later in the monograph.

3.1 Military, Commercial and Humanitarian Logistics:

A Comparison
According to U.S. Field Manual 100-16 [35], “military logistics is the
process of planning and executing the movement and sustainment of
operating forces in the execution of military strategy and operations.”
It is the foundation of combat power — the bridge that connects the
nation’s industrial base to its operating forces. This definition can be
further modified to, “a discipline that encompasses the resources that
are needed to keep the means of the military process (operation) going
in order to achieve its desired outputs (objectives). Logistics includes
planning, managing, tracking and controlling these resources” [83].
As reflected in much of the literature, military logistics is regarded
by many as a quantitative discipline. From this perspective, the
20 Design of Humanitarian Logistics

parameters of logistics include amounts of fuel for ship and aircraft,

ordnance, consumable and non-consumable goods, and time windows
such as force accumulation time, order-to-ship time, scheduling time
[22] and many more. There are, of course, many other important logis-
tics issues that are difficult to quantify such as collaboration, informa-
tion and knowledge management, education, and training. Such issues
are unaccounted for by the quantitative analysis. In the rest of the
section we focus on quantitative aspects of military and humanitarian
logistics.
Many of the challenges in commercial logistics management [18,
pp. 3–6] are also reflected in military and humanitarian logistics. How-
ever, there are some significant differences. For example, in commercial
logistics, the distribution network is reconfigured by management if
they deem it necessary to meet changes in demand patterns. Although
demand patterns and demand locations change only occasionally in
commercial enterprises, they change frequently within the context of
conflict and disaster. Thus, the configuration of a distribution network
may be anchored but the last mile distribution is necessarily dynamic
and uncertain for military and also for humanitarian logistics.
Furthermore, while managing inventory is problematic in the retail
environment, it is even more complex in military logistics due to the dis-
tance between locations of commodities and troops. In case of disaster
response same problem is prevalent. For example, inventory manage-
ment of prophylaxis for a pandemic is major issue faced by the con-
trolling agency, such as the Federal Emergency Management Agency
(FEMA) or the Centers for Disease Control and Prevention (CDC). As
a result of instability and uncertainty, concerns about issues of fleet
management, packing, routing, and delivery of supplies within speci-
fied time windows are compounded during times of disaster and con-
flict. Military and humanitarian logistics have significant commonalities
across many of these issues.
Military (especially in the United States) has typically played an
important role in providing support during disasters. Military has an
underlying reliable structure for command and control. This struc-
ture is critical in case of chaotic situation such as war or disaster.
Humanitarian logistics would benefit from military support for ensuring
 3.2 Supply Chain Considerations 21

effectiveness. Setting aside political and security reasons, the common-

ality of situation in war and disaster in conjunction with structure and
resource acquisition in military logistics may help explain why military
support during disasters is often more effective than other forms of aid
(except for certain commercial organizations such as Wal-Mart).
As per funding, military mission normally dictates accomplishment
despite the cost. This does not mean that budget is not a consideration
but supporting of the mission is preferred over saving money. If the
mission is necessary then the funds are made available. Unfortunately,
this may or may not happen in humanitarian logistics since acquisition
of resources is dependent on donations.

3.2 Supply Chain Considerations

A supply chain in its most basic form encompasses three elements:
supply, demand, and flow — flow being the intermediary between the
other two components. In general, a commercial supply chain supplies a
pre-established, standardized product to customers to meet a relatively
constant and forecasted demand via structured resources and contin-
uous flow. In contrast, at any given time, a response supply chain in
humanitarian logistics supplies a wide range of products and services
fulfilling spurts of demand while sharing the flow and capacity with
other relief items [44].
Many businesses believe that making their supply chains faster and
more cost-effective gives them a competitive edge. However, “Only com-
panies that build supply chains that are agile, adaptive and aligned get
ahead of their rivals” [90]. The supply chain for humanitarian relief
must be agile [115], i.e., be able to respond to abruptly changing sup-
ply and demand, and must also be adaptive to market and strategy
changes. For example, when a disaster strikes, information may not be
immediately available to define accurate requirements for the affected
population — and subsequent information may make the initial prepo-
sitioning strategy inappropriate or obsolete. Throughout this supply
chain, from the source to the donors and finally to the customers
(the affected population), the common goal is relief. All the nodes in
between — producers, distributors, and humanitarian agencies — need
22 Design of Humanitarian Logistics

to be aligned to achieve that goal. Since the supply chain for human-
itarian relief and disaster response exists out of necessity, one might
conclude that agility, adaptability, and alignment [90] are particularly
critical for the humanitarian logistics. Unfortunately, although agree-
ment exists amongst humanitarian logisticians about the need for the
alignment of agents in a supply chain, such alignment is difficult to
achieve when the agents consist of a mix of non-profit organizations,
for-profit businesses, and government entities.
The following are factors common to both commercial supply chains
and humanitarian logistics viewed as a supply chain:
• Supply
• Inventory
• Distribution Network
• Flows
• Lead Time
• Information System
• Customer
• Demand
In addition, factors such as the organization’s objective and perfor-
mance measures, and agility, adaptability, and alignment can further
define the supply chain. We now discuss each of these factors from the
perspective of both types of supply chains.
Supply: To meet customer demands, the supplier provides the prod-
uct in the quantity needed by customer. In a commercial supply chain,
the demand has at least two characteristics: the specification of the
product is usually clear and the demand distribution, if not the fore-
casted demand, is also reasonably known. However, in a humanitarian
logistics (unless it is in response to a slow-onset disaster), the relief
items in demand, though from a small list, are diverse in kind and
quantity. The suppliers of these items, other than the ones with whom
contracts already exist, are also not always known in advance. The sup-
pliers could consist of one or more vendors, donors, or a combination
thereof. “Donations place additional complications on the procurement
process, since it is difficult to define what will come from donors, and
what will have to be secured from vendors” [44]. Furthermore, donors
 3.2 Supply Chain Considerations 23

usually do not have comprehensive information about the relief orga-

nization’s existing inventory. The ambiguity and uncertainty of supply
pose a major challenge for humanitarian logistics.
Inventory: With reasonable predictability of demand and inventory
visibility, inventory control is manageable in commercial supply chains.
Absent these factors, inventory control can be extremely difficult in sup-
ply chains for humanitarian purposes [104]. Moreover, any long deliv-
ery time for supplies during a disaster resulting sometimes in outages
forces high inventory levels. An inventory management policy could be
developed by the humanitarian organization to assure a steady stream
of supplies from warehouses to its points of distribution. Depending
on the type of disaster and the shelf life of supplies, storage may be
either short term or long term. For example, in the 1989 relief efforts in
South Sudan, supplies such as grains had to be held for long periods of
time [15]. But in the case of a pandemic, such as bird flu, the vaccines
have a short shelf life. Perishability of the vaccine and its obsolescence
due to a change in virus strain both contribute toward a short shelf
life. Such supply issues complicate the inventory management system,
particularly in humanitarian logistics.
Distribution Network: In a crisis, when the infrastructure is at risk,
transportation and distribution become a significant issue. In cases of
earthquake or flood, roads may not be traversable and the already
challenging last mile distribution in the supply chain becomes an even
greater hurdle. The location of the disaster may further complicate
supply distribution, as observed in the aftermath of the 2005 Pakistan
earthquake due to geography and politics. The methods and vehicles
of distribution generally available in commercial supply chains become
a luxury in a disaster situation.
Flow: The flow of products in most commercial supply chain is man-
aged so as to create a match between demand and supply. However, due
to the large uncertainty around the acquisition of supplies, unknown
demand, and vulnerable infrastructure, the flow of products in a supply
chain for humanitarian response can be anything but smooth.
Lead Time: In a disaster, the ideal lead time of relief supplies from
the victims’ perspective is zero, but the same disaster can induce longer
lead times for the humanitarian organizations. Here, although there
24 Design of Humanitarian Logistics

exists a need for reliable supply of different products with substan-

tially smaller lead time, due to lack of information on requirements,
the humanitarian relief supply chain may not be able to mimic its com-
mercial counterpart. It is interesting to note that certain commercial
organizations such as Wal-Mart and Waffle-house were very effective
in responding to hurricane Katrina disaster in the United States. The
humanitarian response supply chain can learn from the operational
practices of these commercial organizations.
Information System: Supply chains for humanitarian causes have
several systemic deficiencies, one of which is ineffective leveraging of
information technology [155]. Logistics and supply chain management
in the humanitarian sector is mostly a manual process. Information sys-
tems are an integral part of commercial supply chains and information
management is seen as a fundamental supply chain management strat-
egy. This characteristic is best illustrated by the well-known success
of Wal-Mart, which gained phenomenal efficiencies in its supply chain
through such practices as real-time tracking of inventory levels and
purchases at stores, vendor managed inventory, use of RFID (radio fre-
quency identification), etc. “Lack of existing software that can handle
the dynamism of humanitarian supply chain, limited internet access,
computer shortages and lack of trained staff as well as the inability of
IT staff” [155] lead to inefficiencies in humanitarian supply chains.
Customer: “The issue of customer responsiveness is problematic
when considering the humanitarian supply chain” [115]. The customer
in this supply chain is an elusive player. On one hand, the end user
from the affected population is rarely a part of the supply chain. On
the other hand, the donors who provide funding or supplies must be
convinced by the humanitarian organization that the victims who are
in need are reached [164]. Thus in some sense donors are also customers
in humanitarian logistics. Such a supply chain where the supplier and
the end users are both customers who need to be pleased, presents a
dichotomy that is difficult to manage.
Demand: As noted previously, demand patterns in a commer-
cial supply chain — timeline, requirements, quantity, and location —
are known with reasonable accuracy. But humanitarian organizations
often lack comprehensive demand data of this type. Demand patterns
 3.2 Supply Chain Considerations 25

are estimated, normally after the disaster strikes. The only certainty
regarding this unpredictable demand is that the need is immediate.
Requirements, in terms of product type, are either unknown (such as
with a vaccine where the strain is not known with certainty) or are not
communicated (as in the Pakistan earthquake where women burned
donated tweed jackets for fuel). In addition, based on the severity of
the disaster, demand quantity can be estimated, but may not be mea-
sured, and, demand locations, though known, may not be accessible.
Objective: The objective in a supply chain in humanitarian logistics
is evident — to minimize “loss of life and alleviate suffering” [153]. In
the commercial supply chain the strategic goal is “to produce high qual-
ity products at low cost to maximize profit and achieve high customer
satisfaction” [104]. A commercial supply chain is considered successful
when the bottom line profit has increased. The supply chain for human-
itarian response is deemed successful when it “mitigates the urgent
needs of a population with a sustainable reduction of their vulnerabil-
ity in the shortest amount of time and the least amount of resources”
[164]. In case of a supply chain in humanitarian logistics, although these
two goals cannot be achieved simultaneously, certain tradeoffs between
budget and suffering can be established.
Agility: Quick and economical responses to fluctuations in supply
and demand, i.e., agility, is critical to a successful supply chain [90].
Agility is also defined as “the ability to thrive and prosper in an envi-
ronment of constant and unpredictable change” [102]. In supply chains
for humanitarian response, agility is essential: rapid deployment of crit-
ical supplies and services is of utmost importance in the face of unpre-
dictable demand and uncertain supply. Although both types of supply
chains should be agile, the commercial supply chain is relatively sta-
ble whereas the humanitarian response supply chain is not. Agility in
humanitarian supply chain can be created by prepositioning of assets
(responsiveness), quick response (efficiency), and flexibility in the face
of unknown demand. This indeed is one of the strengths of humanitar-
ian logistics “that business could use to improve their performance and
competitive advantage” [164].
Adaptability: Successful commercial supply chains adapt their struc-
ture and strategies as the market evolves [90]. Without adaptability to
26 Design of Humanitarian Logistics

demand, supply, and structure, the humanitarian supply chain cannot

function. Supply chains for humanitarian causes must be able to set
up, change structure, and modify strategies very quickly. In short, they
must be adaptable.
Alignment: The correct alignment of goals of partners in a human-
itarian logistics is more important than that in its commercial coun-
terpart. Most often, a player in a commercial supply chain is looking
to optimize a local function, possibly at the cost of global optimality.
However, for humanitarian efforts, the effective partnership between
private and humanitarian sectors, military and humanitarian organi-
zations, and local authorities and humanitarian communities are all
essential for delivering relief [158]. Though each stakeholder has multi-
ple goals and priorities, significant inefficiencies arise when incentives
are not aligned. “A supply chain works well if the risks, costs and
rewards of doing business are distributed fairly across the network”
[113]. Although describing a commercial supply chain, this statement
is equally true for a supply chain in humanitarian logistics.
Performance Measure: A link between decision making and perfor-
mance measures is fundamental to success. It is a known fact that mea-
sure of performance can drive behavior. Commercial supply chains have
predominantly used two performance measures, “cost, and combination
of cost and customer responsiveness” [13]. However, for humanitarian
logistics, the term “cost” needs to be defined. Is the cost actual money
spent or budget required or the cost to the economy? Is the cost the
fatalities or suffering of the survivors or both? Cost in humanitarian
logistics is some or all these based on whose perspective it is. In addition
to cost as a performance measure, other measures can be applied. The
typical performance measures of a commercial supply chain, resource,
output, and flexibility are also good surrogates for performance mea-
sures in a supply chain for humanitarian causes [14]. If inadequate uti-
lization of resources results in the termination of donation and funding,
performance can also be measured in terms of resources. Lack of speedy
deliveries shows inefficiencies. So speed could be a performance mea-
sure. If a poor relief effort leads to increased suffering and unnecessary
loss of human life, then the casualties can be used as a performance
measure.
 3.3 Humanitarian Logistics as a Supply Chain 27

Fig. 3.1 Time line of humanitarian supply chain.

3.3 Humanitarian Logistics as a Supply Chain

Process flow in humanitarian logistics can be divided into three stages
along the time line (Figure 3.1): preparedness efforts before the disaster
strikes, response immediately after the disaster strikes, and recovery in
the post-disaster period. The first period is strategic: the disaster has
not occurred but the prepositioning of assets and infrastructure prepa-
rations take place in anticipation of a disaster [132]. Prepositioning
of assets may include the expansion of warehouses, medical facilities,
and temporary shelters, while infrastructure preparation may include
provision of airstrips and ramp space at the airfields.
When a disaster strikes, the response follows: donations and fund-
ing are solicited from donors, supplies are obtained from pre-contracted
vendors. Sometimes the supplies are obtained in advance, especially
during the pre-positioning stages. The supplies received from donors
and supplies purchased from vendors are then transported by vari-
ous means to predetermined locations and distributed by emergency
responders in the affected areas. The complexity of humanitarian logis-
tics can be appreciated when the distribution process through this time
line, along with the factors and characteristics of this supply chain are
taken into account. Finally, the recovery effort is an ongoing process
28 Design of Humanitarian Logistics

which must also include collecting data about lessons learned. Increas-
ingly, to resolve some of the major issues described in this section so
far, humanitarian logisticians and academics are relying on mathemati-
cal modeling, a topic we discuss next. As mentioned earlier, in the next
section, we base our discussion of analytical models concerning supply
chain issues in humanitarian logistics on the three natural phases of
preparation, response, and recovery (relief) operations.
 4
Supply Chain Issues in Humanitarian Logistics

So far we have discussed what humanitarian logistics is. First, we

described the need for it. Based on lessons learned from the previous
disasters we found various inadequacies which give rise to much needed
research. Then we understood it in terms of a classification of different
disasters. We distinguished between disaster response and humanitar-
ian relief and compared it with military and commercial logistics to
frame the supply chain issues. The design of response supply chain nat-
urally leads to the supply chain related issues of humanitarian logistics
which we now discuss in this section.
The motivation behind this discussion is to review the analytical
models as they are applied to humanitarian logistics in preparation,
disaster response and relief, by sampling selected articles from pub-
lished and current research. The discussion in the next two sections
offers the current status of the research which also reveals what has
not been done. This is important for two reasons. One, the academics
working in humanitarian logistics area can now place their research in
the context and second, researchers wanting to work in this area can
find leads to unexplored territories. From this viewpoint, in the last

29
30 Supply Chain Issues in Humanitarian Logistics

section of this monograph we discuss the case studies to uncover cur-

rent practices, thus, divulging the complexity of strategies, their conse-
quences, and the operations in logistics for a humanitarian cause. The
informal survey in the next two sections focuses on certain elements of
the response supply chain and organizational issues of the humanitarian
logistics whereas the case studies offer a wider picture. The academic
research and case studies offer a contrast between the specific focused
issues against the broader and somewhat chaotic situations in an actual
disaster. Case studies are of value especially in evaluating the imple-
mentation of academic research. The case studies in the next section
highlight the research that is needed.
Until recently, emergency planners were focusing on tactical and
even more on operational level of humanitarian logistics. Agility and
adaptability which are vital characteristics of the humanitarian sup-
ply chain cannot be supported without being prepared or preposi-
tioned for adequate capacity and resources. Facility location of the
warehouses for resources plays an important role in this strategic pre-
establishment. History and recent disasters, especially natural disasters
such as Hurricane Katrina in the United States, tsunami in the Indian
Ocean, typhoons in Bangladesh “beg the questions: What assets need
to be in place in anticipation of a disaster? And, where should they
be located?” [132]. After locating the facilities and prepositioning the
assets, inventories of critical supplies must be managed well. Distribu-
tion of these supplies and dispensing of services are equally important
since last mile distribution is an inherent hazard in a supply chain.
Transporting the critical supplies and personnel from source to affected
areas, evacuation of all types of individuals to temporary or permanent
shelters and medical facilities wherever necessary play major roles in
the humanitarian logistics. A decision support system in place encom-
passing all these aspects of the supply chain facilitates reach of the first
responders and avoids fiasco.
From the perspective of the key stages in humanitarian logistics,
preparation, disaster response, and relief, we now proceed to discuss the
supply chain issues that influence the research and action in human-
itarian logistics. Preparation includes problems such as facility loca-
tion, pre-positioning of assets, resource allocation, and planning for
 4.1 Preparation 31

transportation in anticipation of disaster. The actual disaster response

includes problems of inventory management, distribution, and decision
making during the crisis. Amongst these issues, inventory management
is a concern across all these time lines of before, during, and after with
different supplies needed at different time. The issue of relief includes
transportation, evacuation, traffic management, and such.
Some of the approaches mentioned above have been applied to mili-
tary logistics and to commercial supply chain but we focus on problems
from humanitarian logistics angle. Here humanitarian and logistics both
are key words. For example facility location for prepositioning is part of
logistics and emergency services are a substantial part of humanitarian
effort. Therefore, we focus on facility location of emergency services
but not just emergency services.
Though the research reviewed is categorized based on these three
stages of humanitarian logistics, preparation, disaster response, and
ongoing relief, there clearly exists a crossover between these articles.
Locating facilities sometimes also refers to level of inventory and so
does distribution. Distribution also involves transportation and relief
operations include delivery of commodities as well as evacuation. There-
fore this arrangement is more for guidance than categorization.

4.1 Preparation
Preparation in anticipation of a disaster involves prepositioning of
assets and resource allocation. A significant part of it involves loca-
tion of warehouses, distribution centers, and so forth. At the other end
of the spectrum, it involves planning for vehicle replacement. We focus
on these issues in this subsection. Though inventory management, espe-
cially what to store, where to store, and when to store are important
issues in planning they are also part of disaster response. We therefore
discuss inventory management in the next subsection.

4.1.1 Facility Location

Prepositioning of critical supplies and services has been researched
using, predominantly, two types of optimization models, set covering
problem (SCP) and facility location problem (FLP). In a set covering
32 Supply Chain Issues in Humanitarian Logistics

problem each facility is associated with a region and it is strictly a

binary problem asking a question does one build/locate or not. There
is no “how to build/locate” involved in it. The facility location prob-
lem is a network-based problem where one chooses a facility from a
candidate list making the decision which facility to close or open and
sometimes which one to expand. Both the models decide upon location
but address different issues. Traditionally, the latter is known as an
FLP in the context of distribution networks or supply chains. Some of
the research in facility location and prepositioning discussed below is
summarized in terms of major decision and objective of the model, and
the method in Table 4.1.
Emergency planners often must deal with design or reconfiguration
of disaster systems. Number of approaches developed by researchers are
based on optimization problems such as minimum cost maximal cov-
ering problems known as location set covering problem (LSCP). The
objective function typically minimizes the number of facilities required.
LSCP, since then has been extended to problems where instead of dis-
tance, coverage time is used. LSCP also requires covering all demand
points which may not be realistic in face of budgets. Maximal Covering
Location Problem (MCLP) does not require the coverage of all demand
points due to limited budget which is modeled as a constraint of placing
“p” facilities to maximize coverage. In the past, Brandeau and Chiu [19]
provided a broad overview of over 50 representative problems in loca-
tion research. Klose and Drexl [79] describe the current state-of-the-art
facility location models for distribution system design.
Determining good locations of facilities with special attention to
“maximum time or distance that separates a user from his closest ser-
vice”, dates back to the article by Toregas et al. [161]. This model is
most applicable to locations of emergency service facilities such as fire
stations though it can easily represent schools or libraries. There have
been various modifications of MCLP model as reviewed extensively by
Marianov and ReVelle [97] for emergency services. Locating emergency
services is critical in preparing for disasters and there has been signifi-
cant work done dealing with this issue [26, 49, 59, 150, 161].
An interesting example for locating emergency services is of basing
Airtankers for forest fire control in Ontario, Canada [96]. The problem
 4.1 Preparation 33

Table 4.1. Summary of Literature Facility Location and prepositioning of assets.

Article Major decision Objective Method
Balcik and Number and locations of Maximize total Maximal covering
Beamon [10] distributing centers in a expected demand location model
relief network coverage integrating facility
location and
inventory decision
Cataldi et al. Location of distributing Maximize reduction Optimization model,
[24] facilities and in Maleria risk solved with
deployment plan heuristics
Dekle et al. [33] Identify potential disaster Minimize total Covering Location
recovery centers number of Problem with a
disaster relief two-stage approach
centers for each
county residence
within a given
distance
Duran [36] Given initial investment Minimize average Prepositioning model
find the network response time
congiguration
Hale and Establish an efficient Minimize number of Location Set Covering
Moberg [60] network of secure sites supporting Problem
storage facilities multiple supply
chain facilities
Lee et al. [88] Number of dispensing sites Minimize number of Facility location model
points of
dispensing
(POD)
McCall [105] Identify preposition Minimize “victim- Location mixed-integer
locations for pack-up nautical-miles” programming model
kits to transport kits
to each disaster
location
Ozdamar et al. Logistic plan indicating Minimize Hybrid of
[119] optimal mixed pick up unstatisfied multi-commodity
and delivery schedules demand network flow and
along with optimal vehicle routing
quantities and types of problem
loads picked up and
delivered on the routes
Rawls and Location of emergency Minimize expected Two-stage stochastic
Turnquist supplies and allocating costs over all optimization model
[126] quantities of those scenarios influenced by
supplies facility location and
resource allocation
Salmeron and First stage decisions for Minimized expected Two-stage stochastic
Apte [132] expansion of resources number of optimization model
and second stage for casualties and
logistics then expected
number of people
left behind
Tean [151] Prepositioning decisions Maximize expected Two-stage stochasitic
for resources rescued survivors model
and delivery of
commodities
34 Supply Chain Issues in Humanitarian Logistics

of specifying how many airtankers should be home-based and how much

excess capacity should be used to satisfy daily airtanker deployment
needs is modeled as an integer linear program with several simplifying
assumptions. This model was applied regionally by Ontario Ministry
of Natural Resources to place CL-215 airtankers for 1993 fire season.
An important issue in humanitarian logistics, especially in emer-
gency services locations is the definition of objective. An analysis of
private and public sector location models [130] suggests that though
fundamentally the models are similar, the objectives in each sector dif-
fer. In public sector this task is harder due to the difficulty in attaching
monetary value to social utility. In the 1970s extensive research was
done in locating facilities on a network to reach and meet the demands
of customers. Objective of these models evolved from minimizing costs
to maximizing public welfare [129].
Finding “robust facility location” that will not only be well suited
based on the current requirements and the state of the system, but
should continue to be the best sites for the changes in the system in
terms of environment, population density, and market trends may not
be of immediate concern during a crisis but is relevant in the context
of humanitarian logistics. Owen and Daskin [117] address the strategic
nature of facility location problems by stochastic or dynamic charac-
teristics. Dynamic treatment focuses on time perspective for longer
horizon, whereas stochastic treatment focuses on uncertainty in input
parameter such as demand forecast and distances. The stochastic for-
mulations are either based on explicit probability distribution or are
managed with stochasticity by scenario planning.
The location problem for the Florida County Disaster Recovery
Centers is solved by modeling it as an LSCP and with a two-stage
solution approach [33]. It is a great example of what can be achieved if
academics, practitioners, and emergency planners collaborate in solv-
ing the problem. FEMA opens Disaster Recovery Centers (DRC) in the
affected areas. The team of authors helps FEMA identify at least three
(as required by FEMA) DRC sites in Alachua County, Florida, United
States. Their choice of objectives is to minimize average travel distance
to the closest DRC, minimize maximum travel distance to DRC, min-
imize total number of DRCs needed within a specified distance of a
 4.1 Preparation 35

nearest DRC, and maximize probability that at least one DRC will be
useable after the disaster. FEMA and the authors agreed on minimizing
total number of DRCs for three different distances.
They formulate the problem as a set covering problem and solve
the original model using aggregate data with Microsoft Excel by solv-
ing p-center problems in succession. It should be recognized that the
authors accepted the loss of accuracy in reducing the problem size but
demonstrated that the emergency planners could use the same aggrega-
tion method to simplify and increase solvability achieving user friend-
liness. It should be noted that this simplification does not capture the
entire problem but is effective. The whole process provides a means
for testing various scenarios. Various algorithms are picked for aggre-
gation of the demand points. Important take away from this article is
not just location of DRCs but some of the valuable lessons learned in
such projects are that the sponsors (in this case FEMA) may not have
a single overall objective, the choice of model may be subjective, data
collection is most of the work, and last but not the least, simplification
by aggregation of data may be necessary.
An important recommendation made by FEMA is to store critical
supplies in a safe and secure location. “However, storing a set of these
items at every distribution center, manufacturing facility, transporta-
tion hub, and office within the supply chain can be cost prohibitive”,
hence Hale and Moberg [60] propose “secure site selection process”
that “can balance operational effectiveness and cost-efficiency by iden-
tifying the minimum number and possible locations of off-site storage
facilities”. The research combines recommendations from FEMA with
LSCP model.
Using a four-step decision process the authors create a list of poten-
tial sites. The steps involved are identifying the critical resources needed
at each site, identifying all critical sites within the supply chain, set-
ting maximum response time and minimum distance goals, and using
these steps to identify the numbers and approximate location of emer-
gency resource storage facilities. Based on this candidate list an LSCP
is developed to minimize the total number of sites. In addition to limit-
ing the distances from the sites and other facilities of the humanitarian
supply chain the model has constraints to incorporate the fact that the
36 Supply Chain Issues in Humanitarian Logistics

managers must “decide upon a minimum distance each secure location

must be from each supply chain facility”. Together, these constraints
make a solution space smaller and manageable yielding solutions to
this problem relatively quicker than the standard generalized network
LSCP. The model finds three locations for secure sites in the three dark
shaded areas from given list of seven critical facilities.
In case of a bioterrorist attack, “mass dispensing requires the rapid
establishment of a network of dispensing sites and health facilities that
are flexible, scalable and sustainable for medical prophylaxis and treat-
ment of general population” [88]. Authors use facility location models to
determine number of points of dispensing (POD). The model accounts
for population densities, maximum travel distance, list of available pri-
vate and public facilities, and the availability of staff required to operate
PODs.
The approach in this research study has two steps. The first step
determines the minimum number of PODs required in a given region
subject to maximum distance a household has to travel. There is an
additional constraint for requiring at least two PODs to incorporate
the uncertainty of possible closing of one POD due to disaster. This is
a capacitated FLP. This model when solved returns minimum number
of PODs required. The second step minimizes average distance traveled
by households for the given number of sites. In addition to sensitivity
analysis offering insights into various parameters of the situation the
authors compare “heterogeneous mix of PODs” such as public PODs,
drive-through and walk-through PODs, private PODs for large busi-
nesses, university/college campuses, nursing homes for elderly, jails,
and airport. They also point out that dispensing itself is sensitive to
the availability of critical staff to operate the PODs. This research finds
site locations based on diverse but necessary properties of a site. The
extensive and in-depth study is then applied to mass dispensing plan-
ning effort for the Atlanta metropolitan area.
Prevention or intervention in case of disaster outbreaks may be
thought of as a prepositioning strategy that is prevalent in cases such
as Malaria infestation in Africa. Deploying limited resources such as
facilities, labor, insecticide, and bed-nets while maximizing the reduc-
tion of risks due to Malaria with constrained budget [24] is modeled
 4.1 Preparation 37

as an optimization problem and solved using heuristic. The decisions

such as location of distributing centers, their covering zones, deploying
of intervention in these zones, number of people protected, resources,
and their allocation at distribution centers are included in the model.
Rather than solving all parts simultaneously the location decision is
done using heuristics whereas deployment is solved optimally. The
model is applied to the case of Malaria interventions in Africa. The
data are provided by WHO and all models are implemented in Excel
for use by non-government organizations for countries in Africa.
Yushimito et al. [168] address the problem of facility location which
can prove to be useful for prepositioning supplies in the affected area
from a disaster. They formulate an uncapacitated facility location prob-
lem to minimize response time. The heuristic algorithm is based on
Voronoi diagrams.

4.1.2 Logistics Planning

The research discussed so far focuses on facility location which is an
integral part of preparation. However, it is extremely important to cover
all aspects of planning in anticipation of disaster. Humanitarian logis-
tics planning in case of disaster predominantly includes shipping critical
supplies and services to affected areas and evacuate affected population.
An overall model encompassing logistics planning in emergency situ-
ation integrating natural disaster logistics decision support system is
developed by Ozdamar et al. [119]. This research, we believe, is one of
the first of the kind addressing dynamic time-dependent transportation
problem for ongoing aid delivery incorporating new requests for supplies
and availability of modes of transportation. Though very operational in
nature the model is meant for emergency planning. The authors model
the situation as a hybrid of multi-commodity network flow problem and
vehicle routing problem. The model can be decomposed into two multi-
commodity problems, first a linear model for conventional commodities
and second for vehicle flow and hence an integer model. The solution
process uses Lagrangian Relaxation on arc capacity constraints. The
authors test the algorithm on a small set of test instances of an earth-
quake scenario of realistic size.
38 Supply Chain Issues in Humanitarian Logistics

Yi and Ozdamar [167] examine the problem of coordination

of “transportation of commodities from major supply centers to
distribution centers in affected areas and the transport of wounded
people from affected areas to temporary and permanent emergency
unit”. The authors develop a mixed-integer multi-commodity network
flow model treating vehicles as integer commodity flows in the first
stage and providing schedules for the same using “vehicle splitting algo-
rithm”. The objective is to minimize delay in supplying critical com-
modities and health services. The approach is compared with vehicle
routing problem single-stage formulation approach. To validate their
model the authors implemented it on a possible scenario of a severe
earthquake in Istanbul, Turkey using public source data.
The Center for Emergency Response Analytics (CERA) uses
discrete-event simulation to evaluate POD for responding to an
Anthrax attack. POD operations are deemed to be extremely com-
plex since the “interactions among the process layout, staffing levels,
physical limitations of the POD site, client transportation, and sup-
ply chain operations defy analysis by static tools and techniques, such
as spreadsheets and mathematical models” [165]. Developed simula-
tion models facilitate a complete plan for response. This model makes
explicit assumptions such as restricting the attack to an island and that
this was the only attack. The effectiveness of dispensing is evaluated by
measuring the performance of cycle time for a family, waiting time in
traffic, length of car-back-up, client population in the POD, formulation
of queues in the process, and utilization of staff. The end results of the
simulation identify importance of assumptions, whether community’s
goal was achievable, compared alternatives and more importantly eval-
uate how the plan would work in spite of incorrect assumptions. The
simulation model is robust enough to extend to dispensing the second
time around for further treatment.
“Though the location and capacities of resources provided are key
components in the disaster management plan, little research has been
conducted on the topic of prior planning” [126]. The authors there-
fore offer a two-stage stochastic optimization model for determining
location and quantity of critical supplies. They model uncertainty of
demand requirements and locations. They also model the uncertainties
 4.1 Preparation 39

of the transportation network and loss of supplies due to disaster. The

supplies include equipment, consumables, and non-consumable goods,
but no services. The model is formulated with cost minimizing objec-
tive where the costs are the expected costs of supplies resulting from
the supplier locations, procurement, allocation, and transportation.
The model is solved and applied to southeastern United States as a
case study. The case study is based on four major and four less severe
hurricanes with 21 developed scenarios including isolated and combined
storms with assigned probabilities of occurrence. The resulting solution
specifies where storage sites should be located.
The issue of planning for transportation of critical supplies and
services and deployment of emergency personnel is discussed by Bar-
barosoglu and Arda [11]. The authors propose a two-stage stochastic
programming framework in case of an earthquake. A multi-commodity,
multi-modal network is formulated to transport goods and personnel.
The uncertainty of a disaster in terms of timing and severity is
captured by random variables characterizing resource requirements
and resource mobilization. Specifically, randomness is represented by
random arc capacity, supply quantity, and demand. The two-stage
stochastic programming structure of the problem has pre-event and
post-event phases. The post-event phase is further divided into early
response and response.
The objective function minimizes the total first-stage transportation
and expected recourse cost and the constraints are standard capac-
ity, supply, and demand constraints in addition to a recourse con-
straint which is determined by solving the second-stage problem for
each scenario. The model is validated using data from the August 1999
earthquake in Turkey. The authors compare their computational results
with wait-and-see solutions (expected value of the optimal solutions if
the perfect information were available) to find value of the stochastic
solution.

4.1.3 Prepositioning and Resource Allocation

Contribution by graduate military students to recent research work
in humanitarian logistics has been in the area of prepositioning of
40 Supply Chain Issues in Humanitarian Logistics

resources. United States Pacific fleet operation is in the area susceptible

to sudden-onset natural disasters. In addition, United States Military
resources are also in danger of terroristic attacks. The humanitarian
relief operations in these instances can be carried out by United States
Navy due to its unique capabilities. McCall [105] identifies the need of
“pre-positioned pack-up kit of first response material” for immediate
deployment taking the “guesswork out of initial requirements”. Two
types of pack-up kits depending on hot or cold weather are preposi-
tioned. The optimization model prescribes a candidate list of storage
locations based on budget and space constraints and quality of each
type of kit. The objective function minimizes the cost of delivering kits
from prepositioned location to potential disaster site in the formulated
location model under the assumptions that a plan for transportation
and distribution is in place. The solution is found by CPLEX within
an implementation of GAMS. Based on World Health Organization
(WHO) Emergency Disaster Database disaster scenarios are created
by varying severity and type. Sensitivity analysis performed for unlim-
ited to limited capacities and budget provides insights.
Prepositioning in the true sense of the way should not be limited to
critical supplies alone. A two-stage probabilistic model offers guidance
in addition to prepositioning of critical supplies, and assets with phys-
ical, budgetary, and capacity limitations [151]. First-stage decisions
describe the expansion of various assets such as warehouses, medical
facilities, and airfields way before disaster strikes. In the second stage
these assets are utilized to maximize “expected rescued survivors”.
Maximizing delivery of commodities is also part of the objective here.
Second-stage decisions are based on the concept of recourse where sup-
plies and personnel are deployed after the uncertain event of disaster
has occurred and assets have already been allocated. The model is
solved using GAMS and CPLEX.
This two-stage stochastic optimization model [151] is evaluated
using two scenarios [62] affecting Washington D.C. metropolitan area
in the United States. Viable strategic options are analyzed from a Cat-
egory 4 hurricane and a one-kiloton nuclear explosion near the city
center. Though notional, the approach does seem plausible since the
data were gathered using public sources and subject matter experts.
 4.1 Preparation 41

The research also discusses policy issues in the United States of “gap
of pain” defined as measures of time from exhaustion of local and
state resources till the arrival of federal resources, which in itself was
a major issue in hurricane Katrina. The study uncovers a more impor-
tant humanitarian logistics question of “what, and how much do they
need?” as opposed to “how much can we send?”
Salmeron and Apte [132] investigate this problem of strategic alloca-
tion of resources for humanitarian responses to future disasters further.
They develop a two-stage stochastic optimization model to address key
issues in current planning, focus being the pre-establishment of ade-
quate capacity and other resources (decisions which need to be made
well-in-advance before a disaster strikes), that enable efficient relief
operations, with the main objective of minimizing the expected num-
ber of casualties. The research is restricted to sudden cyclical disasters
which include natural events such as floods and hurricanes. The model
guides the allocation of budget to acquire and position relief assets.
The study deals with uncertainty regarding the location and sever-
ity of the possible disaster by considering different scenarios, each of
which may occur with a known probability. Knowing the exact sce-
nario would be better for planning purposes, however, that would not
be realistic. The model includes first-stage (strategic) decisions to rep-
resent the expansion of resources such as warehouses, medical facilities
with personnel, ramp spaces, and shelters. All these decisions must be
made well-in-advance of a disaster event, i.e., with only probabilistic
information about its location and severity. Second-stage (operational)
variables in the model plan operations spanning three days following
the event. These concern the logistics of the problem, where allocated
resources and contracted transportation assets are deployed to rescue
and treat survivors in need of immediate medical assistance, deliver
required commodities, and transport the displaced people, all of which
are scenario-dependent.
Computational results on this notional test case provide insights
into the problem complexity and prove the benefit of using optimization
to guide budget allocation. The results show that matching existing
transportation capacity and health capacity for emergency survivors
appears to be the most critical issue. However, as more funding becomes
42 Supply Chain Issues in Humanitarian Logistics

available, expansion of warehouses and delivery of commodities takes

priority because the cost of additional special transportation and health
facilities for the last pockets of emergency survivors is too expensive.
The research also assesses the benefit of the stochastic model by using
“Wait-and-See” (perfect information model) and “Value of Stochastic
Solution” (averaging scenarios model) analyses. It is found in their test
cases the stochastic model reduces the expected number of casualties
compared to deterministic models that are based on perfect information
and/or average of all scenarios.
A facility location model incorporating inventory decisions devel-
oped for a sudden-onset disaster, determining the number and location
of distribution centers and the quantity of inventory at each distribu-
tion center, is a variant of the maximal covering location model [10].
The objective is to maximize “the total expected demand covered by
the established distribution centers”. The model is applied to 286 sce-
narios, each defined by disaster location and its impact, and 45 candi-
date distribution centers are considered based on data gathered from
National Geophysical Data Center. Rest of the data are hypothetical.
The computational analysis shows how the “before” and “after” disas-
ter funding affects performance of the relief system (response time and
percentage of demand satisfied).
CARE International is one of the largest humanitarian organiza-
tions providing relief to the affected population of a disaster-struck
area. Improving agility for acquiring resources and distributing them is
critical. Collaborative research on this topic is done by CARE and
Center for Humanitarian Logistics at Georgia Tech, United States.
The model developed [36] evaluates the effects of prepositioning relief
supplies to reduce average response time. The model tries to answer
the question, “given an initial investment, what is the configuration of
the network that minimizes the average response time?” The objective
function in the mixed-integer program minimizes the expected average
response time over demand instances. The solution offers provided guid-
ance for CARE’s managers to form the prepositioning network. The
results illustrate how to best allocate existing inventory given an oper-
ating budget. The observed results do not differ significantly between
the historical and simulated scenarios.
 4.1 Preparation 43

A project for CARE-USA [31] applies a global commodity tracking

system to improve CARE’s emergency preparedness. They propose that
this will improve the inventory management and accountability through
monitoring and oversight. They also improve warehousing with design
and layout.
The immediate resource requirements and their temporal patterns
are developed using quantitative methods by Holguin-Veras et al.
[64, 65]. The researchers analyze the data set based on Action Request
Forms issued after Hurricane Katrina made landfall. The data show
only 150 different commodities were requested. This conclusion sug-
gests that emergency responders should focus their efforts on trans-
porting these commodities to reduce cycle time of delivery. In addition,
regional prepositioning of these commodities will give a leg-up not only
on transportation time but also will be extremely cost-effective due to
reduced safety stock.
Holguin-Veras et al. [63] identify three broad categories of issues
after Katrina. First, initial impact of the system through scope and
scale of the needs, effects on communication system. Second, institu-
tional issues such as lack of staff and available but untrained staff,
lack of integration between systems and third, logistical issues such as
inefficiencies in prepositioning, unplanned handling and distribution of
donations, acquisition, and limited asset visibility.
Muckstadt et al. [112] discuss a prototype warehousing system
model for response in an Anthrax attack. They propose a two-stage
stochastic model to estimate the operating cost over time and the delay
in emergency response for a given supply chain. The model determines
inventory levels for stochasticity in timing and nature of the event. The
model is optimized in both the stages when exact details are not known
but the emergency event has taken place.
Most of the pandemic-related research is focused on health care
issues than humanitarian logistics. However, such problems pose unique
sets of challenges in emergency preparedness planning in terms of vac-
cine supply chain. Therefore, we offer a small sample of such articles.
Araz et al. [6] study the increasing concerns about the possible social
and economical disruption may arise due to such disasters. They pro-
pose a geospatial–temporal disease spread simulation model to study
44 Supply Chain Issues in Humanitarian Logistics

the effects of policies such as school closure on communities. Schindler

and Radichel [139] develop an agent-based modeling setup and tool to
allow real-time decision support and resource allocation for personnel of
PODs conducting mass vaccination for pandemic influenza. There are
numerous recent articles describing models especially using simulation.
In preparation for disasters, planning of transportation assets plays
an important role. Though transportation has grown to be the second
largest operating cost for humanitarian operations since 1990s [98, 99]
professional fleet management is almost non-existent in smaller NGOs
[98, 99, 100]. The triple-A framework [90] of agility, adaptability and
alignment is used to analyze an interesting supply chain of vehicles
[147, 146]. The authors in these research studies identify the objective
to be agility. They achieve this by compensating with cost-effectiveness
which is accomplished by sacrificing speed of the vehicles.
A research study [100] investigates the vehicle replacement policy
in the International Committee of the Red Cross, one of the largest
humanitarian organizations. They use logistic models to determine
drivers of replacement and linear regression to determine the drivers of
salvage value. A dynamic programming application compares costs of
replacement versus status quo. Using field data collected, the models
were executed and the results indicate that there will be considerable
savings gain by adjusting the replacement policy.
As discussed earlier, humanitarian logistics to some extent is a
response supply chain. Researchers in the following articles have treated
it like one and explored various aspects of it. This focus varies from
design of the supply chain to information flow in the supply chain to
push–pull boundaries of the supply chain.
Resource allocation is being studied through “informal” modeling
by Gralla et al. [54]. This methodology is inspired by the heuristics that
humanitarian logisticians apply to design their supply chains quickly
but not necessarily optimally. Johnson [73] creates an environmental
sustainability scorecard for MedShare International. Surplus medical
supplies and used equipment are collected by MedShare to redistribute
to undeserved hospitals and clinics. The cost tradeoff includes costs
from environmental impact, carbon footprint, etc. The supply chain is
evaluated based on collection, facilities, shipping, and recipient. Mobile
 4.2 Disaster Response 45

facilities are frequently used to deliver diverse and remote requests for
relief. Halper and Raghavan [61] explore best strategies to route these
facilities to maximize services to the affected areas.
Lack of adequate information flow results in more difficulty in man-
aging a humanitarian supply chain. Day et al. [30] are currently working
on finding issues that hinder this flow of information after a large catas-
trophic disaster has struck. The researcher conducted a case study by
focusing on diverse sources of information including government, profit
and non-profit organizations and individuals during and after the disas-
ter. A lightweight web-based system with central database to track the
available supply of beds, share the demand information, and trace the
historic occupancy, is being designed and implemented by Iyer et al.
[71] on a standard web browser through a simple user interface.
Islam [70] identifies potential bottlenecks in the existing business
processes of the supply chain structure of a leading NGO to stay ahead
of the rapidly growing operations in Africa. The conclusions lead to
issues in current procurement processes, multi-levels of approval pro-
cesses, inflexible policies, etc. Aviles et al. [7] address variability in
the timing and quantity of donation to investigate the issue of supply
chain operations. The problem modeled as a network of the WFP sup-
ply chain, decided the time and location of source and transportation of
food. It is also designed to evaluate borrowing against future donations
to allow for food purchases. Push–pull boundaries in a drug supply
chain in Zambia are being studied by Yadav [166] by exploring effects
of inter-organizational coordination and risk-sharing. This drug flow
supply chain coordinates amongst multiple donors. The research dis-
cusses how “push pull boundaries” can be explained using the forecast
driven and order driven supplies in global health supply chain.

4.2 Disaster Response

Disaster response consists of various logistic operations as we have seen
so far. In this subsection we focus on inventory management and dis-
tribution which are important part of the response supply chain. Due
to the significance of these topics they have their own place in man-
aging the supply chain issues of humanitarian logistics. However, it is
46 Supply Chain Issues in Humanitarian Logistics

important to point out that decision making is also a critical issue in

disaster response and hence is discussed in this subsection.

4.2.1 Inventory Management

As every researcher in the field of operations management and opera-
tions research knows extensive research has been (and is being) done
on the topic of Inventory Management, and a large number of research
articles have been written on the topic. However, a literature survey
revealed that very few studies have dealt directly with the management
of inventory in humanitarian relief or disaster response. In dealing with
a disaster, due to unreliability of transportation infrastructure, inven-
tory management becomes all the more critical and hence this topic
needs attention from researchers. We believe, the body of research car-
ried out without the emergency context can be adapted, to some extent,
for managing inventory in humanitarian operations in areas such as
warehousing of critical supplies, determining inventory levels in distri-
bution centers, and stockpiling prophylaxis in case of bioterrorism or
a pandemic. Summary of some of the articles reviewed in inventory
management is given in Table 4.2.
The unpredictable consequences of the disasters, and due to
unknown characteristics, their uncertain remedies give rise to irregu-
lar supply and demand patterns. In addition, there exist issues that are
inherent in disaster situations which “surpass the capabilities of current

Table 4.2. Summary of literature review on inventory management.

Article Major decision Objective Method

Beamon and Optimal order quantities Optimize inventory Stochastic inventory
Kotleba [15] and reorder points for management control model
long-term emergency
relief response
Bravata et al. Stockpile of prophylaxis Evaluate costs and Supply chain model
[21] and dispensing benefits of based on four
alternative strategies
strategies for
inventory of
prophylaxis
Lodree and Proactive disaster Optimize inventory Stochastic inventory
Taskin [95] recovery planning management control model
formulated with
Bayesian updates
 4.2 Disaster Response 47

emergency approaches” [16]. The authors address one of the limitations

of the current literature on inventory management in humanitarian
logistics by “developing a stochastic inventory control model that deter-
mines optimal order quantities and reorder points for a long term
emergency relief response”.
The authors develop this model for responding to crisis in Sudan.
The multi-supplier inventory model develops an inventory policy tak-
ing into consideration the unique demand patterns over a long-term
complex humanitarian relief response. Their models are extensions of
continuous review inventory models with two options for resupply [110].
Beamon and Kotleba [16] develop this model using two reorder levels
(r1 and r2 ) and two-order quantities (Q1 and Q2 ) one for normal sup-
plies and the other for emergency supplies. They assume r1 > r2 and
lead time τ1 of the first order to be greater than τ2 , the lead time of the
emergency reorder. The model gives higher fixed and per unit ordering
costs to emergency reorders than the first order. The authors assume
that international supplier does not run out of supplies, an order quan-
tity is large enough to cover the first reorder so as not to trigger an
additional reorder (Q1 > r1 ), second lead time for emergency orders
and demand occurring is discrete 10-day interval. In addition, for sim-
plification, they assume replenishment lead time for normal orders to
be constant. The model is solved using five-step procedure. Data for
the implementation were collected on-site at Lockichoggio, Kenya. The
research is one of the earlier strategic steps in developing inventory
levels for humanitarian relief.
An application of three single-item inventory management
strategies to humanitarian relief operations in South Sudan [15]
“demonstrates the performance benefits of using quantitative methods
to manage inventory in a relief setting”. First strategy is based on the
model discussed above [16], second is a “heuristic humanitarian relief
model” using simulation. This approach identifies factors critical to
the performance of the system and also confirms that it was adaptable
and could be easily modified to incorporate changes, strategic or
operational. This model demonstrates consequences of backorders.
Third strategy based on the “naive” model was simplistic but was not
as effective as the other two.
48 Supply Chain Issues in Humanitarian Logistics

In most circumstances, a relief worker is not in a position to com-

pute optimal solutions due to lack of technology and time [15, 155].
In addition, due to high turnover of staff, volunteer participation and
limited funding, inventory management policies need to be “easy to
implement, flexible to change and require little effort to maintain”.
Therefore, a heuristic may be more suitable in such circumstances than
an exact algorithm. The heuristics proposed in the second strategy,
called “Silver-Meal heuristic” [141], selects replenishment quantity so
that the total cost per unit time for the duration of replenishment is
minimized. The approach used in this heuristics does not guarantee a
global optimal but guarantees a local minimum for holding and set-up
costs. Based on the implementation and analysis, this strategy with a
lower cost accomplishes greater flexibility. The drawback of this strat-
egy is that under certain scenario it reorders large quantities making
storage difficult especially in the relief period.
Following articles explore interesting aspects of inventory manage-
ment such as non-traditional models, assumptions that demand is a
function of hurricane characteristics and tradeoffs between modes of
transportation. One of the fundamental issues for humanitarian logis-
ticians responding to bioterrorism is the inventory of prophylaxis. It is
therefore important to analyze the strategies that will maintain and dis-
pense local and regional supplies of antibiotics and medical items [21].
The authors do not use a mathematical model but the situation is mod-
eled by considering the local and regional supply chain as well as local
capacity based on four strategies. The strategies are, enhance bioterror-
ism event detection, increase local dispensing capacity, increase local
inventories, and increase deployed inventory. The authors conclude that
mortality is highly dependent on the local dispensing capacity, the
number of individuals requiring prophylaxis, adherence to prophylactic
antibiotics, and delays in attack detection. If regional inventories can
be made available, the bottleneck is local dispensing capacity. Hence it
is more beneficial to increase such dispensing capacity rather than to
increase inventory.
McCoy [108] develops an inventory management model for investi-
gating two tradeoffs, expensive airlift of a product for speedy delivery
versus less expensive but slower shipment (and thus the ability to ship
 4.2 Disaster Response 49

more total material), and acquisition and management of an expensive

stockpile versus more budget allocated to quick response. A mathe-
matical approach to Triage is being developed by Uzun et al. [163] to
identify effective dynamic triage rules to maximize total lives saved.
In case of natural disasters such as hurricanes, the prepositioning
of assets and inventory levels at warehouses is complicated due to the
uncertainty in the number of hurricanes expected to make landfall.
The humanitarian relief organizations rely on pre-contracted vendors
and retailers for critical supplies like batteries, flashlights, generators,
etc. Therefore, it is important to analyze the problem from the perspec-
tive of manufacturers or retailers [95]. The statistical model is formu-
lated “as an optimal stopping problem with Bayesian updates, where
updates are based on hurricane predictions”. The authors assume that
hurricane-related demand is a function of hurricane characteristics such
as its intensity and path, and they further assume that demand is a
function of storm’s maximum sustained wind speed. The model consid-
ers demands associated with “regular” hurricane as well as “extreme”
hurricane. The distribution for random variable that specifies whether
the storm is extreme is Bernoulli. The order (or production cost) is
an increasing sequence for forecast updates. The approach is based on
the concept from emergency management that “risk identification is an
integral part of the planning process”. The model is implemented for
data from HURDAT database, which is a database maintained by the
National Hurricane Center of attributes of hurricanes since 1851 in the
Atlantic Basin.
A comprehensive data is analyzed to confirm a common belief that
in an economic downturn food banks face difficulties to serve the needy
[57]. Of interest here is the way inventory is evaluated, not by weight
but by nutritional value. The researchers develop a model for one-
commodity pick up and delivery vehicle routing problem and employ
various methods to solve the problem.

4.2.2 Distribution
In a disaster response or humanitarian relief mission the emergency
planners encounter problems of transporting critical supplies from the
50 Supply Chain Issues in Humanitarian Logistics

Table 4.3. Summary of literature review on distribution.

Article Major decision Objective Method

Ekici et al. [39] Food distribution Maximize distribution Disease spread
logistics during of food to victims model integrated
pandemic and of pandemic with facility
intervention strategies location and
resource
allocation
Ergun et al. [43] Choose facilities Minimize travel time Network model with
and facility congestion games
congestion using Nash
equilibrium for
polynomial time
Feng and Keller Qualitatively and Establish distribution Multiple-objective
[46] quantitatively evaluate plan that satisfies decision analysis
distribution plans both local and
national
governments
objectives
Khan and Determine alternate Optimize dispensing Multi-criterian
Richter [75] modes of dispensing in of prophylaxis decision analysis
Dept. of Health, Los
Angeles County
Lee et al. Plan large-scale Minimize total cost in Resource allocation
[86, 87] emergency dispensing terms of cycle time, model and
throughput, and simulation
actual cost at POD
Liang et al. [94] Analyze error disaster Improve information Fault-tolerant
avoidance mechanism distribution system design
reliability with several
layers of
hierarchical setup

source to the affected areas within a given time frame. Though this
can be handled as a prepositioning strategy, the problem can also be
modeled as a distribution problem. Some of the research articles in
distribution are summarized in Table 4.3.
Rathi et al. [125] formulate the problem of distribution with three
linear programming formulations. First two formulations minimize the
tardiness of supplies delivered within the specified time window. The
models do this by determining which transportation assets can be opti-
mally allocated in each time period. The third model distributes the
lateness evenly over all routes. The formulation of this type offers
“tradeoffs between the degree of control to be exercised by the planner
 4.2 Disaster Response 51

and the speed of computation”. The models were implemented at

United States Transportation Command within the Deployment Anal-
ysis Prototype System.
Emergency planners work together with the Center for Disease Con-
trol (CDC) for a response to bioterrorism or pandemic by means of
planning mass dispensing of prophylaxis. In addition to locating the
right POD at right place, Lee et al. [86, 87] develop a decision sup-
port system. The system combines optimization tools with simulation
techniques. The relative fast execution of the system shows that “a
real-time decision support system is viable through careful design of a
stand-alone simulator coupled with powerful tailor-designed optimiza-
tion solvers”. This extensive study brings flexibility to “what if” scenar-
ios of the decision support system. The distribution of medications can
be conveniently altered proving it to be an invaluable tool, not only in
mass dispensing, but in training the planners and strategically prepar-
ing for disaster response. The speed and the flexibility also contribute
to “dynamic on-the-fly reconfiguration of large PODs.”
The authors [86, 87] point out that though there do exist simi-
larities between POD network and manufacturing and service system
design there also are some fundamental differences such as estimates
of parameters and distribution due to uncertainty of human behavior
during receiving or dispensing of medication in a disaster situation. The
POD site is modeled as an open queuing network where the nodes of a
directed graph represent stations in the process. The staffing problem is
modeled as an optimization problem. The objective is to minimize the
total cost, measured in terms of cycle time, throughput, and cost at each
station. This is subject to, in addition to integrality constraints, num-
ber of workers allocated to each station being within an interval and
availability of workers. The model developed is a mixed-integer program
with non-linear objective function. Simulation assists the optimization
in resolving and updating the resource allocation statistics. The system
was validated as a field study with 864,000 households by replicating
the smallpox tests performed by [103].
One of the effective safety measures common to thyroid cancer,
resulting from exposure to radioactive iodine in a nuclear incident,
is potassium iodide (KI). Hence, the distribution plans for KI are
52 Supply Chain Issues in Humanitarian Logistics

necessary in preparing for such disasters [46]. The authors develop a

multiple-objective decision analysis tool to “qualitatively and quan-
titatively” evaluate the plans for a hypothetical region. The decision
makers in such situations, government agencies both nationwide and
local, must be in agreement about the objective. A hierarchy of objec-
tives can help them make a better decision. The multiple objectives are
converted to single-attribute value function by adding the values with
swing weights to solve the problem. The “what if” analysis is made
possible by choosing different weights for different regions (“sliders”).
Research study by Sheu [140] presents a hybrid fuzzy clustering-
optimization approach to the distribution of critical supplies in the
affected areas during the rescue period. Based on five assumptions
related to the affected areas and their geographic location — the sup-
ply sources, relief distribution centers and channels, disaster-induced
damage conditions and casualties, time varying relief demand, and
types of relief supplies — the model is developed using the follow-
ing steps: time-varying relief demand forecast, grouping of the affected
areas (this involves binary transformation for fuzzy clustering), deter-
mination of distribution priority, group-based relief distribution, and
finally, dynamic relief supply. The research validates the approach based
on the massive earthquake in Taiwan in September 1999. The numer-
ical results reported show that the performance can be improved to
30.6% using this approach.
Last mile distribution is hard enough in commercial supply chain
but due to the unique challenges it is even harder in disaster response
and in humanitarian relief. Reaching the affected areas from local dis-
tribution centers (LDCs) is discussed by Balcik et al. [10] by developing
an optimization model. The principal decisions are allocation of crit-
ical supplies to LDCs and delivery scheduling of vehicles. Demand is
divided in to two groups, Type 1 and Type 2. Type 1 demand (tools,
blankets, etc.) occurs once at the start of the planning horizon. Type 2
demand consists of consumable goods occurs periodically. The first type
is large in quantity whereas the second type may not be so and for this
no backorders exist. The mixed-integer programming model minimizes
the sum of routing and penalty costs. The approach has two phases. In
the first phase all possible delivery routes are generated. A candidate
 4.2 Disaster Response 53

list is extracted from these routes by preprocessing for consideration the

infrastructure changes or the availability of vehicles. Based on the list
of routes, periods to visit each demand location, the amount of supplies
and type, and the delivery routes are identified by executing the model
in the second phase. The authors validate this approach by implement-
ing the model using numerical examples. The example helps illustrate
relationship between decision variables and parameters. Solution times
are long for larger problems.
Distribution of critical supplies may be accomplished if the affected
population can reach the supplies as opposed to supplies reaching the
individuals. These individuals are “decentralized decision makers who
choose to visit one of a number of facilities opened by a centralized
planner” [43]. Choice of the facility to visit depends on the individual
and that behavior impacts the system of self-help. Instances of such
distribution may be during a pandemic or a slow-onset disaster.
The model being developed in this research can have two objec-
tives, one standard objective of minimizing the travel time and the
other, more expanded version that includes facility congestion. Due to
this objective the problem becomes asymmetric network configuration
game. The network is defined with costs on arcs as delays experienced
by the users. The flow on the network is “unweighted, unsplittable,
and atomic”. The model examines the problem of mutual assignment
with individuals wanting to minimize their own cost and the centralized
planner trying to minimize the system-wide cost. Equilibrium solution
for the first objective is found easily with tight bounds on the prices of
“anarchy and stability”. Nash equilibrium for the second objective is
found in polynomial time. The authors conclude that using only opti-
mization approach leads to poor system performance for decentralized
decision making. Therefore, they integrate that approach with game
theory to develop these models.
Lessons learned from past incidents of pandemic cases suggest
that efficient and quick response can reduce morbidity, mortality, and
costs to the community. In collaboration with American Red Cross,
Ekici et al. [39] develop a “disease spread model” integrating facility
location and resource allocation network model for food distribution.
They also develop a heuristics to find “near-optimal solutions in large
54 Supply Chain Issues in Humanitarian Logistics

instances”. First an individual-based continuous time stochastic model

for influenza transmission is constructed comprising the progress of the
disease within an infected person and the spread amongst the affected
population with extension in the form of intervention such as voluntary
quarantine or school closure. Then they develop a “capacitated multi-
period hierarchical facility location problem (CMPH-FLP)”. Since this
is an NP-hard problem, which is difficult to solve with a commercial
solvers, an “Add-Drop” heuristic is also developed. Using these models
and solution approach they propose resolution for location of major
facilities and PODs, closing of these sites based on need for food and
minimum cost to serve target population by allocating food. For the
food distribution network a deterministic as well as dynamic approach
was compared with “Perfect Solution” benchmark case. Authors report
that deterministic models could be used in advance planning whereas
dynamic models are more appropriate for response. The approach was
validated by implementing the models in the state of Georgia.
Another research study for choosing modes of dispensing prophy-
laxis based on multi-criterion decision analysis is done and implemented
in Los Angeles County by Khan and Richter [74, 75]. The authors
follow six steps, identifying stakeholders, deciding on objective hierar-
chy based on relevant measures, identifying value functions, identifying
weight of relative importance, assessing the alternatives, and conduct-
ing sensitive analysis in the context of mass dispensing of prophylaxis.
This research study including the model (especially the transportation
model) and analyses helped the Department of Health at Los Ange-
les County to gain support from their plan in alternative modes of
dispensing. It also provided insight for the emergency planners. The
authors point out however, that “the process can be repeated by any
jurisdiction but definition of “best” will rely on issues and gaps that
are identified with the jurisdiction’s POD plans for mass prophylaxis.
Apte and Ferrer [4] identify the essential infrastructure needed to
develop a cold-chain distribution network for vaccine in a military appli-
cation. They develop a network distribution model for transportation
of elements needed for successful vaccine flow through the network.
Various characteristics of the supply chain are identified, including a
discussion on metrics of distribution centers, support for these and the
 4.2 Disaster Response 55

inventory necessary at these centers. In addition, they follow the infor-

mation and vaccine flow through the distribution network.
Griffin [55] studies efficiencies and effectiveness of drug distribution
in Zambia. The study first follows forecasting and then creates map
of transportation. The researcher proposes changes to supply chain
structure and distribution strategies. Since the current system fails due
to duplication, costs overrun, inadequate supply, and uncertainties of
donations. Menezes and Varela [109] analyze real requirement of medi-
cal and non-food items for Medicos Sin Fronteras. They also investigate
optimal distribution policies by comparing air and/or sea routes. The
researchers use Zimbabwe for case study.
Distribution of blood by monitoring collection and usage patterns
is being studied with validation through a case study in Zambia by
Osuntogun et al. [116]. The forecasting model is based on probabilistic
model of individual units of blood. Flow assignment of blood from
collection centers to transfusion outlets with fair and efficient utilization
of resources.
Distribution of vaccines and prophylaxis in the event of a substan-
tial public health emergency is a necessity that requires planning on
part of state or local health departments. King and Muckstadt [76]
create numerous mathematical models focusing on the entire response
network of strategic national stockpile of Center for Disease Control.
Some of the logistics questions they pose and develop the models for
are about the location of warehouses [112], policies of locating the facil-
ities and their effects on costs and resource requirements, optimal levels
of inventory under unbounded capacity to balance inventory costs and
backorder penalties, capacities of all the resources in the response sup-
ply chain such as staff at PODs, optimal levels of service, and allocation
of additional investment for improving services.
In addition to actual distribution of critical physical supplies and
medical services, there exist some other important factors in distri-
bution. For instance, it is also essential that individuals receive the
updated contents of the software when accessed. A research study [94]
“analyzes the error situations of a proposed content-delivering service
that has been implemented in a large-scaled company in Taiwan”.
Disasters can cause computers to crash and thus interrupt transmission
56 Supply Chain Issues in Humanitarian Logistics

of data which results in hindering of emergency information distri-

bution. The authors propose “disaster avoidance mechanism” based
on reliability model and waiting-time distribution of disconnectivity.
They describe a fault-tolerant design of the system which they then
implement within a large service company with several layers in a
hierarchical setup. The study claims that such an approach, though
very expensive, is “clearly” more reliable than a six-sigma continuous
improvement process.
The delivery of critical supplies and services to the affected area
gives rise to a convergence process since personnel, information, and
material flow toward the affected area. Holguin-Veras and Destro
[63] focus on quantifying this material convergence by assembling a
database derived from Katrina donations. They investigate whether
donations can be explained in terms of the severity of the disaster or
by socioeconomic properties of the geographical location of the event
and donors. Such research can help in the management of donations in
the response process.

4.2.3 Decision Making

Decision making in case of an emergency, whether natural or manmade,
is difficult due to various uncertainties and the lack of information. Yet
it must be carried out quickly and correctly. Developing decision sup-
port systems is a fundamental aspect of humanitarian logistics. Some
of the articles in decision making are summarized in Table 4.4.
Dai et al. [28] study design and development of “emergency deci-
sion support systems (EDSS)”. They compare the “conventional deci-
sion making (CDM)” with “emergency decision making (EDM)”. They
develop a computerized safety protection and disaster response for
coal mines in China. The models discussed are “ventilation network
model, escape evacuation model and rescue materials demand model”.
The ventilation network model takes into account a set of non-linear
algebraic equations incorporating various environmental issues. The
escape evacuation model is a multichannel quickest path problem with
weighted network. And rescue materials problem are a set of “expe-
rience formulas”. The authors report that using a distributed system
 4.2 Disaster Response 57

Table 4.4. Summary of literature on decision making.

Article Major decision Objective Method

Ak et al. [1] Demand forecasting, Minimize costs and Forecasting,
transportation cost maximize transportation
estimation, bid award service to cost estimation
allocation recipients and bid award
allocation models
Dai et al. [28] Design and development Compare Non-linear algebraic
of computerized conventional equations and
support systems for decision making multichannel
emergency decision with emergency quickest path
making decision making model with
weighted network.
Engineer et al. Disease prevention/ Optimize Dynamic
[42] immunizing vaccination programming
vaccination schedule schedules for
for children childhood
immunization
Regnier and Evaluate trade-off Minimize cost (or Discrete Markov
Harr [128] between lead time and fatalities) chain model
forecast accuracy

helps users share data, models, and knowledge, but more importantly
they are able to utilize subject matter experts.
In anticipation of a disaster, humanitarian relief or disaster response
is predominantly posed as a cost (or fatalities) minimizing problem
sometimes using stochasticity for location and severity. However, “the
value of waiting for updated forecasts” is “neglected” [128]. The authors
develop a decision model where the decision maker expects and plans
for updated forecasts. The accuracy of forecast increases as lead time for
preparation decreases. They model the situation as a discrete Markov
chain. The value derived from expected forecasts depends on the
“specific alternatives and cost profile” of the decision maker. An imple-
mentation to areas such as Norfolk, Virginia and Galveston, Texas, in
the United States show a conceptual saving of up to 8% when compared
to iterative static solutions.
Decision making in Public Health and Medicine is increasingly
made based on mathematical and simulation models in case of
disaster response. Brandeau et al. [20] review these models and offer
recommendation for best practices for developing such model. This is
a survey of all such models and a position paper. The authors propose
58 Supply Chain Issues in Humanitarian Logistics

that the response model address real-world problems, be user friendly,

cover complex issues in a simple manner, include relevant outcomes,
and incorporate uncertainties. The authors conclude that such models
are critical for effective planning for responding to disaster, especially
in the health sector and their recommendation “can increase the
applicability and interpretability of future models, thereby improving
strategic, tactical and operational aspects of preparedness planning
and response” [20].
The World Health Organization (WHO), through the Pan Ameri-
can Health Organization (PAHO), acquires vaccines for 37 countries in
Latin America and the Caribbean [1]. Currently a team of researchers
from PAHO and Georgia Tech are developing an analytical tool kit
to make decisions about demand forecasting, transportation costs, and
bid awards in order to improve the supply chain for vaccines.
Carbajal et al. [23] are conducting research to understand
which analytical methodology leads to better decisions in Debris
Management.

4.3 Relief Operations

Relief operations belong at the end of the response supply chain in
continuation of providing relief. These constitute on transportation of
critical supplies and evacuation of the affected. Both of these rely heav-
ily on robustness of the infrastructure. However, a menace to both these
issues is disruption of traffic.

4.3.1 Transportation
As mentioned in the earlier subsection, Rathi et al. [125], offered three
models for distribution as well as transportation of critical supplies for
carrying out the relief operations within given time windows. Haghani
and Oh [58] develop a large-scale multi-commodity, multi-modal net-
work flow problem with time windows to transport a variety of critical
supplies using a vehicle fleet from sources to affected areas in time, effec-
tively and efficiently. The model can be an important part of decision-
making tool for emergency planner. They convert a physical network
 4.3 Relief Operations 59

to time–space network to incorporate the dynamic decision process.

The traffic on the physical network is divided into three types. First
(routing) type of traffic moves on from one node to another node on a
certain type of vehicle, second (transfer) type of traffic changes mode
of transportation at a node and the third type carries the supply or
demand over to the next time period. The assumptions are that trans-
fer is allowed at transshipment nodes which include origin nodes that
are also transshipment nodes and all costs are linear. The objective
function minimizes the sum of “the vehicular flow costs, the commod-
ity flow costs, the supply or demand carry-over costs and transfer costs
over all time periods”. The three sets of constraints, include commod-
ity flow constraints, vehicular flow constraints, and linkage constraints
between vehicles and commodity flow.
The authors report two solution approaches, one decomposes the
model into sub-problems by relaxing linkage constraints and the other
fixes integer variables, at every iteration. They implement the model
on artificially generated data set on a “medium-sized physical net-
work”. The empirical study designed to analyze robustness suggests
that the model is sensitive to the arc capacity, vehicle capacity, vehic-
ular flow costs and demand carry-over costs, moderately sensitive in
certain range of demand and supply, and very sensitive beyond that. It
is also moderately sensitive to commodity flow costs and supply carry-
over costs and insensitive to transfer costs.
In a disaster situation, especially earthquake and flooding, when
roads are not traversable, helicopter missions are common. The emer-
gency personnel or decision makers need an efficient and effective pro-
cedure to deploy the suitable crew/fleet configuration or appropriate
flight routes. Barbarosoglu et al. [12] develop mathematical models for
such situations. The models execute the operational and tactical mis-
sions of,

“D1: the determination of the helicopter fleet composition by

assigning helicopters from the air force bases to the oper-
ation base.
D2: the assignment of pilots with given aviation capabilities to
the helicopters.
60 Supply Chain Issues in Humanitarian Logistics

D3: the determination of the number of tours to be undertaken

by each helicopter.
D4: the vehicle routing of helicopters from the operation base to
disaster points in the emergency area.
D5: the load/unload, delivery, transshipment and rescue plans of
each helicopter in every tour.
D6: the re-fueling schedule of each helicopter at the operation
base.”

The decomposed model at the top level makes a decision on D1, D2,
and D3 whereas at the base level it makes a decision on D4, D5, and
D6. The objective in the first case minimizes the cost of selecting dif-
ferent helicopters and pilots from the air force bases whereas at the
base level it minimizes maximum tour duration among all helicopter
which include refueling times. The solution approach is an iterative
process that exchanges information between the tactical problem and
operational problem which the authors call “Hierarchical multi-criteria
approach”. The approach is implemented on databased on scenarios
similar to disaster relief data from the Turkish Army. The scenarios
are kept dimensionally small due to computational complexity of the
problem. The contribution of the study lies in the result that a deci-
sion maker, based on his preference and aspirations can choose from
“non-dominated” solutions.
DeAngelis et al. [32] incorporate the travel of airplanes between
depots and destinations and parking of the airplane during the nights
for World Food Program (WFP) in Angola. The problem can be
described by depots (each one with limited supply and parking space),
planes having the same capacity for carrying supplies but having vary-
ing speed (can park at any depot subject to availability) and commu-
nities where food is being delivered (food measured in plane-load and
time measured in days). The depots are identified in five types based
on where a plane parks the previous day and present day. The objec-
tive function maximizes the total satisfied demand subject to supply–
demand constraints, and fairly standard vehicle routing constraints.
One fundamental constraint that is included is based on classification of
 4.3 Relief Operations 61

depots. Effectiveness of the model is checked using four test cases based
on WFP Angola records for May 2001. The implementation improved
number of trips by 9–23%. The authors believe that the weekly sched-
ules produced by the model are a good starting point for decision
makers.
Table 4.5 summarizing research done so far in transportation.

Table 4.5. Summary of literature on transportation.

Article Major decision Objective Method

Balcik et al. Allocate relief supplies at Minimize Mixed-integer
[10] local distribution transportation programming
centers and determine cost and model
delivery maximize the
schedule/routes benefits to
recipients
Barbarosoglu Determine number of Minimize cost of Hierarchical
et al. [12] helicopters, number of selecting multi-criteria and
tours, their routing helicopters/pilots two-level
and delivery schedules and maximum decomposition
tour duration
among all
helicopters
Barbarosoglu Transportation of critical Minimize Two-stage
and Arda supplies and services transportation stochastic model;
[11] and recourse cost multi-commodity,
multi-modal
network
DeAngelis et al. Plan for travel between Maximize total Integer linear
[32] depots and clients, satisfied demand programming
deliver and park at model
depot
Eksioglu et al. Determine routes for Maximize system Optimization and
[40] vehicles without mobility simulation
communication means models
in a disaster
Haghani and Transportation of Minimize sum of Large-scale
Oh [58] supplies and relief vehicular flow multi-commodity,
personnel costs, commodity multi-modal
flow costs, the network
supply or demand
carry-over costs,
and transfer costs
Sheu [140] Distribution of relief Optimize Hybrid fuzzy
supplies distribution of clustering-
critical supplies in optimization
affected areas
62 Supply Chain Issues in Humanitarian Logistics

4.3.2 Evacuation
As mentioned earlier, evacuation is a significant part of relief opera-
tions. In disaster situations government and non-government officials
arrange to evacuate people with or without their own transportation.
This type of evacuation involves various issues. Type of disaster will
dictate total or partial evacuation to distant or near-by relief location
and whether it is long term or temporary. Condition of the infrastruc-
ture will force which mode of transport to use but the evacuation also
depends on the available fleet of vehicles and their capacity. Circum-
stances surrounding the evacuee, whether they are able or not, phys-
ically, financially, or emotionally also play major roles. Summary of
some of the research in evacuation in Table 4.6 is given.

Table 4.6. Summary of literature on evacuation.

Article Major decision Objective Method

Bakuli and Resize passageways Improve throughput Queuing network
Smith [8] and total egress model
time
Chiu and Zheng Simultaneous Minimize travel time Linear programming
[25] mobilization over entire system model
destination, traffic
assignment, and
departure schedule
Regnier [127] Reduced decision lead Relationship Stochastic model of
times from 72 to 48 between decision storm movement
hours making lead time and statistical
and Atlantic hurricane track
hurricane model
uncertainty
Regnier and Evacuation Explore relationship Simulation and
Harr [128] between lead time analysis
and track
uncertainty
Sayyady and Evacuating affected Minimize number of Mixed-integer linear
Eksioglu population dependent casualties and program and
[138] on public transit total evac time Tabu search
algorithm
Yi and Logistics planning for Minimize unsatisfied Mixed-integer,
Ozdamar dispatching demand and multi-commodity
[167] commodities and waiting of the network flow
evacuation of wounded wounded (service model
delay)
 4.3 Relief Operations 63

One of the major decisions in evacuation in case of a disaster is the

confinement of the structure from which one needs to evacuate. Bakuli
and Smith [8] investigate the problem of designing emergency evacua-
tion networks by extending the state-dependent queuing network mod-
els by incorporating the mean value analysis algorithm within Powell’s
derivative free constrained optimization algorithm. They discuss the
effect of change in circulation widths on throughput and implement
the design to resize passageways which in turn improve “both through-
put and total egress time”.
A more focused issue of evacuating “multi-priority groups (MPG)”
is studied by Chiu and Zheng [25]. For suddenonset disasters they
develop a Linear Programming model for multiple emergency responses
and evacuation of MPG to different destinations with varying priority.
The authors assume that guidance for evacuation is done ideally, the
population of evacuees is known, the area from which evacuation is to
take place is known, and the “complete evacuation scenario is of inter-
est and there is no shadow evacuation”. Under these assumptions the
objective is to minimize travel time over the entire system based on
priorities. The problem is then transformed from algebraic formulation
to Matrix form for facilitation of obtaining solution in a user-friendly
platform. The approach was validated on a test network of 8 nodes and
14 directed arcs. The authors believe that this study is a starting point
for “modeling a complex evacuation logistics mobilization problem”.
In addition to actual evacuation, a major issue in a disaster (espe-
cially hurricane) response for evacuation decisions is that the forecasts
provide imperfect information. A research study [127] uses a stochastic
model of movement of storm to explore relationship between the lead
time for decision making and uncertainty of tracking the Atlantic hur-
ricanes. A statistical hurricane track model [128] is used for strike prob-
abilities. The author reports the results of this relationship as reducing
decision lead times from 72 hours to 48 hours could reduce evacuation
costs. The article also examines the cost of false alarms.
Emergency planners frequently depend on public transit system for
evacuation in the case of a suddenonset disaster, especially in the urban
areas. Sayyady and Eksioglu [138] propose a methodology that would
64 Supply Chain Issues in Humanitarian Logistics

be beneficial for evacuating affected population that depends on public

transit for mobility. The proposed model is a mixed-integer linear pro-
gramming problem where the objective function minimizes the num-
ber of casualties and total evacuation time. The prototype problem
is solved using Tabu search algorithm on a numerical example with
generated data. To validate the approach authors implemented it to
the south-central part of the city of Fort Worth, Texas, United States.
They also conduct simulation study to examine the quality of solutions
obtained by Tabu search.
During Hurricane Katrina in United States the evacuation plan
required the affected people to meet up with buses at select locations in
order to be evacuated. Unfortunately, this type of plan does not help
those people who are unable to move themselves to the designated
meeting locations. Post-Katrina Emergency Management Reform Act
in the United States of 2006 states that state or local governments
have the responsibility to coordinate evacuation plans for all popula-
tions, including those with disabilities. Apte and Heath [5] develop an
optimization model to build an executable routing in order to dispatch
multiple vehicles to pick up disabled persons from their individual loca-
tions and evacuate them. They plan to test the model by identifying a
geographic location, and testing the model with sample data.
A research study [40] incorporates an information center that uses
prediction and optimization models and heuristic algorithms to gener-
ate alternate routes due to disruption in traffic and infrastructure that
are results of a disaster. The proposed framework was tested using
simulation model on a generated scenario. The results show that the
proposed framework increases the mobility and average speed.
Minimizing the effects of disruption in traffic flow due to disasters,
natural or manmade, is critical for various reasons. Jin and Eksioglu
[72] develop a mathematical model to reduce the delay by incorporating
a prediction model for vehicles without communication capabilities or
for vehicles that use a communication tool called QUALCOM. An algo-
rithm is also proposed to update the parameters of the mathematical
model. The model and the algorithm are being tested using simulated
disaster scenario. The results so far suggest that the proposed models
improve mobility and average speed.
 4.3 Relief Operations 65

In this section, we discussed the factors that influence the effective-

ness and efficiency of humanitarian supply chain, such as prepositioning
and facility location, inventory management, distribution, transporta-
tion, evacuation, and decision making, in view of preparation, response,
and relief. However, in addition to the operational issues organizational
issues are also critical. The focus of the next section is on organizational
factors that affect humanitarian logistics.
 5
Organizational Issues in Humanitarian Logistics

The organizational issues that will facilitate humanitarian logistics

include several factors that are strategic. We focus on some of the
issues listed as follows:

• Collaboration among players

• Training and education of emergency staff
• Role of donors and their donations
• Information and knowledge management
• Risk management.

The order here does not necessarily indicate the relative importance of
the factors. For example, according to Tomasini and Van Wassenhove
[158] information and knowledge management are two most critical
issues whereas approaching humanitarian logistics from risk manage-
ment perspective [77, 85, 158] is considered important for assessing
disruption to supply chains [9, 17, 47, 48, 50, 51, 91, 89, 131, 145].

5.1 Collaboration
Humanitarian organizations offer stability during disaster through
relief operations. Gibbons [53] raises questions about collaboration

66
 5.1 Collaboration 67

among emergency planners, community leaders, and non-profit agen-

cies through organizational issues. The author identifies attributes of
disaster response that may increase effectiveness of operation. Collabo-
ration also needs to happen at all level [121]. Especially, the population
affected by the disaster must be a part of the solution for relief and be
engaged in the recovery effort.
Examining the past experience reveals the critical need for collabo-
ration and coordination [158] among not only the humanitarian organi-
zations but also among the military, private and humanitarian sectors
[124, 134, 164]. One of the primary collaboration, especially for trans-
portation and distribution of critical supplies and services and security
of these activities, is between military and humanitarian agencies. “The
use of military forces in emergency relief can offer benefits to the inter-
national relief system, although such involvement is controversial due
to a number of practical, political and ethical issues” [124]. Although
military is strong in logistical and organizational structure, due to its
primary goal is fighting the wars, this goal is perceived to be in some-
what of a conflict with the goal of humanitarian assistance.
Having said the above, there have been numerous instances, includ-
ing during Hurricane Katrina, where division of United States Marine
Corps (USMC), “Combat Logistics Battalion (CLB) for Marine Expe-
ditionary Unit has the inherent mission of providing initial support to
humanitarian assistance (HA) operations” [149]. Military assets pro-
vide unique capabilities not found in the civilian humanitarian response
system [101]. United Nations Office of Coordination of Humanitar-
ian Affairs (OCHA) believes that proximity and understanding of the
neighboring countries’ political, social, and geographic environment is
the reason why military assets are usually the first ones to be deployed.
However, there remain challenges in this civil–military collaboration in
terms of principles versus pragmatism and lack of military doctrine and
training in humanitarian relief operations.
Collaboration between public (humanitarian) agencies and private
sector is also a necessity in humanitarian supply chain, especially in the
last mile distribution [157]. However, due to their inherently opposing
characteristics, whereby humanitarian organizations can be slow and
bureaucratic while “the other a fast-moving action-oriented business”,
68 Organizational Issues in Humanitarian Logistics

collaboration can become difficult [164]. Yadav [166] describes how

collaboration amongst multiple donors can improve the global supply
chain for medicines. Such examples illustrate the need and the type
of collaboration and the success of the same among various players. It
should be noted that such collaboration critically depends on sharing
information and taking actions based on lessons learned.
Building relationships is critical in humanitarian logistics. Eggen-
hofer et al. [37] explore design options for crisis response manage-
ment and its information and communication technology. Sotnikov [144]
studies effects of partnerships of local health departments on prepared-
ness in case of a disaster. Statistical significance between the correla-
tions quantifies the findings.

5.2 Information and Knowledge Management

Information Management is critical to the success or failure of human-
itarian logistics [101]. In addition, as discussed earlier, a humanitar-
ian supply chain needs to be agile. It can only achieve agility through
an effective information infrastructure and knowledge of requirements
in the affected area [90, 115, 158]. A humanitarian supply chain can
gauge and respond to the demand, if, the gathering of information does
not pose a formidable challenge. As was found in the first two earth-
quakes in El Salvador in 2001 the information assessment of “what was
affected, what was needed, what resources were available, and as the aid
began to arrive spontaneously, what was coming and when” was nec-
essary for responding quickly [159]. Information may become available
in phases and hence it needs to be managed accordingly — first, when
the event happens; second, when requirements of the affected areas are
known; third, when deployment begins and what is sent is known; and
fourth, when it is determined whether the right kind of supplies and
services have been deployed. A well-managed information system in a
humanitarian supply chain is characterized by having visibility, trans-
parency, and accountability [158]. Visibility into the flow of supplies and
services, transparency of the process of transportation and distribution
that can identify issues in efficiency, and accountability of the players
within the supply chain, form three pillars of information management.
 5.3 Training and Education 69

Every time a disaster strikes, lessons are learned by all involved.

A humanitarian organization collaborates and coordinates during that
disaster and in addition to field experience it also learns about the cul-
tures of all the players involved including the affected population. As
Tomasini and Van Wassenhove [158] point out, knowledge is created
and is needed at different levels. In the affected areas the knowledge
resides with the people, at the supply chain level it resides in the pro-
cess, and it also resides at the contextual level. Untrained staff and
turnover among emergency planners and responders (due to high level
of stress) mandate that knowledge management must belong at the
core of a humanitarian supply chain.
One of the lessons learned from Hurricane Katrina is that people
with chronic illness or disabilities face unique and significant challenges.
Kulemeka [84] is studying to examine how U. S. and Canadian cities
are using the web to inform how to manage such challenges especially
during disaster.

5.3 Training and Education

There exists a general consensus that untrained or uneducated person-
nel may hinder the efficiency and effectiveness of humanitarian logistics.
One of the five fundamental keys to being prepared is human resources
[158]. Humanitarian organizations often face the problem of inadequate
staffing due to various reasons such as high turnover and requirement of
diverse skill-set. Moreover, frequently this staff is neither well-trained
nor well selected (especially when it comes to volunteer staff). But this
pool of personnel must be adequately trained and educated in vari-
ous aspects of humanitarian logistics due to the unique set of activi-
ties and environment they must operate in. This points to the need of
standardization of training in emergency planning and field operations.
The establishment of such standards would “help ensure comparability,
quality assurance and international compatibility of training” [2].

5.4 Role of Donors and Donations

The main difficulty with donors is that they are suppliers as well as
customers. They are suppliers since they donate the commodities or
70 Organizational Issues in Humanitarian Logistics

monies for the commodities. But at the same time they only donate
with generosity provided they are convinced that donations are reach-
ing the needy. As a party interested in assessing performance they are
also customers of this supply chain. In addition, it is difficult to fore-
cast the type and quantity of donations that may become available.
There is high uncertainty about what is going to be provided and what
condition it is going to be in. Unsuitable donations as in recent Pak-
istan earthquake [111], or genetically modified food in Southern Africa
[160], or expired medication in Armenian earthquake in 1988 [44], are
examples of situations where product mix and timing of the donations
were not particularly appropriate. Such issues add to the complexity of
managing humanitarian logistics.

5.5 Risk Management

Disaster preparedness strategies can be developed using a risk manage-
ment framework. Paton [122] describes the process of how “incident,
operational and organizational demands interact with resilience and
vulnerability factors to affect stress risk during mobilization, response
and reintegration phases of disaster response”. In addition to this per-
spective, disruption of humanitarian supply chains can also be under-
stood based on the fundamentals of vulnerabilities of supply chain
[3, 77, 81, 93, 123].
Two categories of risk affect the design and management of a supply
chain; risks due to the lack of coordination between supply and demand,
and risks due to a disruption in the supply chain because of natural
disasters, economic downturns, and terrorists [77]. The first category,
where there are uncertainties of supply as well as demand, was dis-
cussed in Section 2.2 (“Supply Chain Perspective”). In the second cat-
egory where disruption is due to natural disasters the risk management
approaches lead to alternative supply chain models. This is described
by Papadakis and Ziemba [120] where the authors trace the financial
consequences to alternate models of build-to-order as opposed to build-
to-stock. Due to global outsourcing and the lead time involved in such
supply chains, disasters add to the complexity [78]. Kleindorfer and
Saad [77] recommend “Specifying sources of risk and vulnerabilities,
 5.5 Risk Management 71

Assessment and Mitigation” as three fundamental tasks that must be

“practiced continuously and concurrently” based on four premises of
specification of underlying hazard for the risk, quantification of risk
through risk assessment process, attention to decision environment, and
integration of policies and actions with continuous risk assessment and
coordination among supply chain players. These organizational issues
can be better understood if we review the case studies. As mentioned
earlier through the current practices of case studies we get a better per-
spective of inadequacies in the current academic research. Additionally,
in the next section we also describe the challenges faced by the humani-
tarian logisticians. Thus the current practices described and challenges
faced lead us to suggest the need for future research.
 6
Conclusions and Potential for Future Research

In the last two sections we reviewed and discussed research in supply

chain issues based on analytical models and we also perused organiza-
tional issues in humanitarian logistics. Each of these research articles
dealt with important aspects of improving operations during response
and relief. One fact that we need to remember is that to maintain or
increase the efficiency and effectiveness of the humanitarian logistics,
specifics are important but only in the context of the humanitarian
space [158]. In this section we first describe some case studies which
lead us to challenges faced by current humanitarian agencies. These
help us to set the possible agenda for future research. We also outline
in this section.

6.1 Examples of Actual Practice

In this section we describe case studies of some recent disasters. These
case studies uncover the processes underlying humanitarian logistics,
some of the lessons learned, and describe progress made in the last
few years by humanitarian organizations in responding to disasters.
But equally importantly they also direct us toward the needs of the
future research. Analytical models simplify the real-world situations to

72
 6.1 Examples of Actual Practice 73

characterize dynamically complex systems, quantifying the tangible as

well as intangible impacts of certain events and identifying the likely
outcome [66]. However, the case studies offer a contrast to these by
describing the actual practice. We use these case studies of the past
and the challenges of today as experienced by the humanitarian officials
to set an agenda for research in the field of humanitarian logistics by
inspiring the readers to pursue research in this area.
In addition to other lessons learned through these case studies
most of the cases suggest the importance of humanitarian space [158]
governed by three equally important principles, humanity, neutrality,
and impartiality. Humanity implies “that human suffering should be
relieved wherever found”. Neutrality implies “that relief should be pro-
vided without bias or affiliation to a party in the conflict” and impar-
tiality implies that “assistance should be provided without discrimina-
tion and with priority given to the most urgent needs”.
To understand this aspect of humanitarian logistics in the context
of the analytical models and organizational issues discussed in the pre-
vious sections consider the following situation. For example, an emer-
gency official needs to evacuate affected population due to a natural
disaster from a region that covers two areas, first one is smaller but
closer from the temporary collection site and the second is larger but
further away from this site. The first area is not as dire in need of evac-
uation as the second area. First area also has fairly affluent population
whereas the second area is economically depressed.
The question is if the official decides to be in the humanitarian space
with his limited resources and in face of forecasted further deteriora-
tion, which area should he focus on? An analytical model with cost-
minimization objective may result in an optimal humanitarian policy of
evacuating the first group, it being closer. But this would compromise
neutrality and impartiality. Discussion of such questions is critical since
a traditional optimization model may address the operational issues but
not necessarily the supporting principles. These concerns, in general,
have not been adequately addressed in the existing research in the area
of humanitarian logistics.
A severe earthquake of magnitude 7.9 on the Richter scale occurred
in Gujarat, India in January 2001. International Federation of Red
74 Conclusions and Potential for Future Research

Cross and Red Crescent Societies (IFRC) were the “choreographer” of

disaster management. The role of collaboration was critical based on
the necessities of the supply chain such as diverse commodities, glob-
alization, and the dynamic nature and the short time span of its exis-
tence. IFRC was able to extend help to the affected population in this
area using a flexible logistic system, in spite of extreme uncertainties
and limited authority. This successful relief effort occurred despite ini-
tial lack of information and a distinct shortage of resources throughout
the effort [136]. However, this case study uncovers the value of reliable
information of the casualties which is necessary to set up the response
supply chain.
The need for organization is uncovered in the case study done after
the 1999 Armenian earthquake in Colombia [98, 99]. Due to coffee pro-
duction, and Colombia being the major global provider of coffee, this
had lasting effect not only on the home country but also on the whole
world.
A sequel to this case study [137] discusses the efforts made by IFRC
using their failure to respond promptly to the 1998 Hurricane Mitch
disaster that swept through Central America. The case describes five
different resources — people, knowledge, goods, money and humani-
tarian communities — that are essential factors in being prepared for
a disaster.
The genesis of United Nations Joint Logistics Centre (UNJLC)
lies in response to the 1996–1997 crises in eastern Zaire. The concept
evolved until the Afghanistan crisis in 2001 [134]. The case relates how
the process of deployment works and highlights the deficiencies as well
as the strengths of UNJLC’s concept. UNJLC also dealt with oper-
ational and logistical bottlenecks during the first six months of the
Afghanistan crisis in 2001 [135]. The bottlenecks were transportation
(land as well as air), inventory, infrastructure network, and security
for all the activities. But collaboration among humanitarian organiza-
tions and peace-keeping military forces along with operational issues
are also critical to success of response and relief. Continuation of the
response by UNJLC during the second year of the Afghanistan cri-
sis led to strategic nation-building initiatives for planning for such
crises [133].
 6.1 Examples of Actual Practice 75

In the 2010 earthquake, in Haiti, planes could not leave the airport
due to lack of available fuel. They occupied the ground space and for
this reason planes carrying the emergency supplies could not land ham-
pering the disaster aid. Fuel is a critical resource in case of disaster.
The UNJLC played an important role in fuel availability and supply
during the 2003 Iraq crisis. A Fuel Cell is a structure that addresses
the availability and sourcing of fuel to the humanitarian community.
The case study [92] uncovered the need for a structured process for
designing and managing the fuel supply chain. The case points out the
importance of prepositioning of this critical yet overlooked commodity,
and suggests how a supply chain for such a commodity can be managed.
This is an important research area to explore.
Multiple loads of relief supplies arrive usually in the aftermath of
a disaster. Humanitarian Supply Management System, SUMA1 is a
tool for managing such relief supplies from the time the donor pledges
are made to the time supplies reach the affected population. Tomasini
and Van Wassenhove [159] describe the process of implementing SUMA
during the earthquakes in El Salvador in 2001. They discuss how the
humanitarian operation can be complex due to interactions between
various parties in a politicized environment. The issue of humanitarian
space is of note here.
The 2002 Southern Africa food crisis in the context of distribution
of genetically altered food donations leads to various issues that need to
be investigated. The partnering of logistic company with a humanitar-
ian organization for eliminating world hunger points out certain aspects
of humanitarian space. Malawi, Lesotho, and Zimbabwe suffered a food
shortage from environmental, political, demographic, and economic fac-
tors. Much of the food donated was genetically modified, so the food
was refused because it was believed by the local population that the
lasting impact would be worse than the existing shortages. WFP did
its best with the situation, trying to stimulate the local economy and
adjust planning to improve the situation. In addition to other lessons
learned the case study [160] illustrates the need for a humanitarian
relief supply chain to be agile and adaptable.

1 (http://www.disaster-info.net/SUMA/english/WhatisSUMA.htm)
76 Conclusions and Potential for Future Research

The response and relief situations described so far suggest that

resource allocation, bottleneck identification within various services
such as transportation, reliable infrastructure networks, designing and
managing a fuel supply chain, inventory management of food are top-
ics of interest in humanitarian logistics. In addition to the supply chain
issues, organizational issues such as collaboration, information manage-
ment need to be studied to manage the crises better.
A Hastily Formed Network (HFN) is a formalization of the idea of
“quickly forming a team for particular, urgent task, and then disband-
ing it when done” [34]. An HFN is based on following fundamentals:
the network must be established quickly and needs to have players
from different communities working together in a “conversation space”
so that they can plan and execute the action plan in agreement. In this
way, these players are able to fulfill a large and urgent mission [67]. In
disasters, an HFN is useful when the advanced networking (electronic)
technology can help resolve humanitarian agency issues.
Steckler et al. [148] describe a historical synopsis through a case
study of Naval Postgraduate School’s (Monterey, United States) deploy-
ment of HFN immediately after Hurricane Katrina in the Gulf Coast
of Mississippi. The authors provide a step-by-step account of lessons
learned and make recommendations for deployment of communications
infrastructure for disasters in the future. The case study describes in
detail what the network architecture was, how the nodes providing
voice and data connectivity to the first responders as well as to the
local health facilities, government offices, police, and fire station were
located. They also report how this network evolved into PODs and
victim shelters. Such research is needed to find innovative solutions to
peculiar situations.
A similar case study [67] using HFN describes the United States
Naval Station mission to the Caribbean and Central/South America.
The study describes how a highly experienced team of ICT experts
formed HFN to enable real-time collaboration over the entire area for
effective humanitarian relief operations. This case study has a focus
that would be of interest and use to those concerned with military
support of security, stabilization, transition, and reconstruction in face
of a disaster. These cases related to hastily formed networks suggest
 6.2 Challenges Today in Humanitarian Logistics 77

another area where research will lead to interesting solutions to the

issues in disaster response.
Humanitarian logistics in the context of information technology and
collaboration [45] among organizations using information communica-
tion technology (ICT) is found to be an important and critical issue
in organizing a response supply chain. The efficiency of humanitar-
ian organizations can be managed using information technology in the
context of IFRC [80]. The authors explain the information flow in the
humanitarian supply chain which encompasses the range of activities
such as preparedness, assessment through performance evaluation of
the information chain.
Farmer and Johnson [45] discuss ICT as a tool for coordination of
“relief and development”. “NetHope” organization which was founded
by information officers of various organizations, humanitarian organi-
zations such as Save the Children, and private companies such as Cisco
Systems and Microsoft is cited as an example from different strate-
gic views such as “technology finding hub, software services provider,
equipment innovator, procurement service” (provider), networker and
educator or “thought leader”.

6.2 Challenges Today in Humanitarian Logistics

The survey of published research and research currently underway (as
known to the author) has been given as an overview of the type of
analytical techniques and tools and organizational framework useful in
humanitarian logistics. Case studies have described the on-location and
in-field issues of different disaster situations. Interviews with and pre-
sentations of individuals involved in disaster response and humanitar-
ian relief have resulted in the following discussion about the challenges
faced by and need for research in humanitarian logistics.
Academics can play an important role in addressing these needs to
help improve efficiency and effectiveness of preparedness, response, and
recovery operations in case of a disaster. Lessons learned thus far from
previous disasters have led to considerable thought and analysis among
humanitarian logisticians. The issues discussed below are certainly not
78 Conclusions and Potential for Future Research

the only challenges faced by this sector; however, the focus here is from
the perspective of operations management and operations research.
After hurricane Katrina, FEMA has established a Logistics
Directorate to uncover and then meet those uncovered challenges. The
challenges they face are across all the stages of planning, management,
and sustainment [143]. They believe that the standardization of systems
would improve the relief structure. Moreover, FEMA would benefit
from a “comprehensive planning/systems approach to addressing all
hazards supply chain management issues of which quantitative analy-
sis is but one tool” [142].
American Red Cross (ARC) faces numerous challenges in their
logistics system [121] in the process of responding to natural disasters,
preparing communities through education for safety in health, deliver-
ing blood, and many such services. The logistics team, more often, ends
up reacting as opposed to following a preset process. Due to the fact
that every disaster is different, demand varies based on type of disaster.
In a situation of multiple disasters there is a possibility of competition
of demand. A challenge in the response supply chain is “mixed sourcing
of resources”. Another issue in their supply chain is the lack of visibil-
ity of supplies once they leave the warehouse. Cradle-to-grave visibility
needs to be a requirement for all supplies, services, and last but not the
least, costs. An important challenge for the material flow of this supply
chain is damages and lost physical infrastructure such as roads and
communications. Both these agencies, FEMA and ARC, will benefit
from a systemic supply chain analysis of response.
Interaction and preferably collaboration between humanitarian
organization like ARC and FEMA affects the disaster response [106].
State or local governments also depend on United States Northern
Command (NORTHCOM) for assistance and support along the same
principles. More research pointing out the compatibility of these orga-
nizations, additionally with military, is needed.
ARC focuses on three issues due to their importance in the
humanitarian logistics deployment: building capacity at the local level,
planning and offering assistance during recovery, and drawing on
the knowledge from lessons learned during planning and evacuation.
 6.2 Challenges Today in Humanitarian Logistics 79

Managing personnel capacity and assistance from local communities

give rise to challenges in assigning right people to right tasks at right
time. This can be resolved using well-known modeling techniques.
Another critical planning issue is sheltering. This is a “difficult
beast” as characterized by ARC. In addition to the recent research
in prepositioning of supplies, research in planning for shelters is needed
to identify the location and capacities of the shelters.
International Medical Corps provides vital health care services
training for helping people affected by war, natural disasters, and dis-
eases help themselves [68]. The constant need for humanitarian logis-
tics means that long-term development for humanitarian aid, which
also translates to more prepositioning and planning for humanitarian
logistics, is a necessity. The relief efforts must have greatest impact at
the right place and right time. For this the information flow in supply
chains must be robust. A priority should be where the data from the
preferred supplier are captured, the information is translated into that
agency’s terms and definitions, and then “e-follow” it through subcon-
tractors (if they exist) to find delivery information. Different agencies
involved in humanitarian logistics at different times locate their ware-
houses and store their supplies. This results in separate warehouses
and separate PODs. There needs to be consolidation, at least, of the
database of all supplies available on location. Hyde [68] suggests that
there is a need for one universal software package that is capable of
tracking all available supplies in case of disaster to offer transparency
of information and usability to all parties concerned. This is quite a
challenge but it is possible with research for achieving this goal.
Daoudi [29] of WFP, a one-stop-shop with virtual stockpile of
food (a database explaining inventory of supplies), believes that there
needs to be greater “predictability, accountability and partnership” to
improve efficiency and effectiveness of humanitarian response. These
three aspects can be improved, provided there exist benchmarks and
performance measures. At CARE-USA, Gazashvili [52] believes that
measuring performance in terms of “timeliness of response”, “qual-
ity and accountability”, and “competence in core sectors” is essential.
Defining and establishing such metrics of performance measures and
setting benchmarks will add rigor to the lessons learned. It will also
80 Conclusions and Potential for Future Research

provide knowledge and analysis of what worked when. All these trans-
late to the need for more research in collaboration, reliability of food
network, reduction of response time, and performance measure for the
response supply chain.
At the Bureau for Global Health, USAID, Emrey [41] believes that
there exists a need for strong public sector pharmaceutical supply chain
that is managed well. Current supply chain systems are multiple and
hence complex with diverse sources of supply. This is a challenge for
managing inventory and distribution which subsequently dictates more
research. Another complication in this supply is counterfeit medicines,
which create issues such as contamination and disruption in a sup-
ply chain though there exists some research in this field, more is
needed.
Listening to emergency planners and first responders and discussing
their issues and difficulties bring many questions to a researcher’s mind.
For example,

• Unpredictability of the timing of most of the disasters

leads to
◦ How can a state of preparedness be maintained?
◦ How should the adequate inventory be sustained?
◦ How can information about inventory be shared?
◦ Is prepositioning the answer?
• High uncertainty in location leads to
◦ Should partnerships with private sector, that have
outfits located all over in the disaster prone areas, be
pre-negotiated?
◦ What information is available about access and
infrastructure?
◦ Can economies of scale and scope help mitigate the
challenge?
◦ How can infrastructure be dealt with for last mile
distribution?
 6.2 Challenges Today in Humanitarian Logistics 81

• Variability of demand in timing as well as quantity during a

disaster raises questions such as
◦ How can structure for demand be modeled perhaps
based on historical data?
◦ Will transparency and visibility of information within
organization eliminate this problem?
• Unpredictable supply lines in a response chain lead to
◦ Is virtual stockpile the answer?
◦ Where to source, locally or internationally?
◦ How should the donations be handled?
∗ What about modified or altered food?
∗ How about expired medications?
◦ What will guarantee smooth supply of critical sup-
plies and services?
∗ Is funding available for that?
∗ Are contracts in place?
◦ What is adequate? What is fair?
• Speedy response is critical in case of crisis, hence
◦ How can cycle time for relief be reduced?
◦ How can critical transportation infrastructure be
maintained?
◦ How can access to affected areas impact the speed?
◦ What can be done to reduce effect of the location on
last mile distribution?
• Prepositioning of assets may be the key to overcome the first
few challenges in anticipation and then actual disaster. In
that case,
◦ What should be the strategic as well as operational
prepositioning?
◦ What part of budget should be invested in preposi-
tioning as opposed to reactive response?
82 Conclusions and Potential for Future Research

• Objectives of humanitarian supply chain is focused on

response and relief, yet
◦ How does one place monetary value on human suffer-
ing or life?
◦ What should be the performance measures of con-
firming whether the objective has been achieved?
◦ What can be done to achieve a convergence of goals
of all the parties?
• Collaboration and partnership is a key to successful human-
itarian supply chain. In order to accomplish this
◦ Why and how should participating sectors prepare
for disaster?
◦ How can a private sector company, with its main
objective of making money, justify involvement in
humanitarian logistics?
◦ How can competition for the same assets be avoided?
◦ How can local communities be involved to help
understand cultural, geographical, and political
environment?
• One of the major lesson learned from recent history is that
humanitarian aid is going to be perpetual and persistent in
the future. Hence,
◦ Can skills be developed for humanitarian operations
through training and education?
◦ Which sectors would benefit the most from such
program?
◦ What type of program should be designed?

The above-mentioned issues as well as the challenges faced

by various humanitarian logisticians lead to fertile research field
researchers will have to build and develop roadways into it with the
help of emergency planners and responders. In the next subsection
 6.3 Future Research 83

some such opportunities are discussed. Academics know that there is

no such thing as a silver bullet or one single methodology, database or
software that will solve most of the problems in humanitarian logistics.
As Owens [118] put it, “one size does not fit all but it is possible to
find a tool from ONE toolbox”.

6.3 Future Research

Successful humanitarian logistics involve using “analytical skills to
develop and apply tools and techniques to problems in order to struc-
ture complex messes” Van Wassenhove [164]. Humanitarian logistics
are inherently chaotic and, as discussed before, extremely complex.
The complexity can be dealt with if the researchers help meet these
challenges through analytical models and systems bearing in mind the
unique challenges of this field.
One of the complexities is the location of critical facilities to
handle the response or relief. Optimal locations of warehouses and
distribution centers for emergency supplies and services, in addition
to optimal location of help centers in an effort to evacuate, are some
of the challenges in any humanitarian relief or disaster response
instances. Such location problems are part of prepositioning of assets
and are subject to allocation of resources. There exists a need for
research in various aspects of these vital issues. Existing models need
to be enhanced and extended. For example, Salmeron and Apte [132]
model can be enhanced to incorporate alternate objectives such as the
total budget for a desired level of performance of other goals. This
concept is universal in commercial logistics and would be significant
to incorporate it in humanitarian logistics. Being aware of effects
of public policy, ethics and economics is of course necessary here.
Another example is the model developed by Cataldi et al. [24] applying
system-based approach to allocate resources in Malaria prevention as
provided to WHO. This research could be extended to other diseases
in other geographic areas bringing critical issues, such as eliminating
guesswork and unreliability in deployment of prophylaxis, by adding
different sets of constraints and objectives. Such research can be
applicable in location of PODs as discussed by Lee et al. [88].
84 Conclusions and Potential for Future Research

Inventory management is a critical issue in all phases of commercial,

military, or humanitarian logistics. From the perspective of preparation
in the response, Balcik and Beamon [10] offer a model that pro-
vides optimal facility locations for the response supply chain in case
of sudden-onset disaster in terms of number of distribution centers
and amount of stock of emergency supplies. Numerous extensions of
this model can be studied. For example, the assumption made by the
authors in offering inventory policy for a single disaster as opposed to
multiple occurring simultaneously can be removed and models could
be developed for such scenarios.
Another example is where Lodree and Taskin [95] describe a stochas-
tic inventory control for proactive disaster recovery planning. Their
model tackles a hurricane event. But this can be integrated for instance,
with a sophisticated hurricane prediction model such as one described
by Regnier [127]. We believe that stochastic models that incorporate
inventory levels for different types of supplies with different lead times
and reorder points that share capacities of warehouses or distribution
centers in case of disasters of sudden or slow-onset, would benefit the
emergency planners substantially.
Research on performance measures of the response supply chain can
be found in Beamon [13, 14]. However, it would be useful to extend
these concepts further. For example, it would be beneficial to know
how the behavior in the humanitarian sector is being driven by which
performance measures.
In planning or executing transportation of emergency supplies and
service vehicles to the affected population as well as evacuation of the
same population, the infrastructure demands close scrutiny. Lack of
transportation assets and equally important robust traffic management
can hinder extensively in such effort. Though there exist some body
of research in transportation and evacuation using multi-commodity,
multi-criterion, stochastic network models, these models can render
reality if factors, such as public policy and ethics in evacuation of the
disabled, and human assessment and knowledge in transportation, are
incorporated.
For these existing models, in their enhancements and extensions,
there are three extremely important factors that researchers need to
 6.3 Future Research 85

keep in mind. First factor is the data used to validate the model. Due
to data limitations researchers are forced to estimate certain param-
eters and input data. Therefore, the models need to be robust to the
available data. Frequently, data estimates are done in consultation with
emergency planners or persons knowledgeable in the area it would sup-
port validity of such implementation and add to the analysis if actual
data can be used. Benefits could be gained if such database is created
and updated based on disasters events.
Though incorporating “real” data are essential, incorporating real-
ity of the event leads to complex models. Therefore, we believe that
the second factor is solution approach for these models. This issue is
twofold. First, the complexity of models usually means the use of com-
mercially available optimization software. This increases computational
burden and we believe, takes the attention away from managerial con-
sideration. Computational analysis is a vital part of some other fields
but humanitarian logistics will benefit more if the focus of the research
is on developing models and solving them to “near optimality” and not
necessarily to optimality. Therefore, future research should be directed
to develop heuristic approaches as well. Secondly, the solution approach
that uses sophisticated mainframe-based software may not be as user
friendly as desktop-based, easily accessible, and well-known software
package such as, for example, Excel. Research being user friendly is crit-
ical to its implementation. Hence, we believe developing such solution
approaches is important if the research is to be of use to the humani-
tarian logistics community.
All this analytical approach needs to take place in the humanitar-
ian space as discussed earlier. Research in this newly defined area of
humanity, neutrality, and impartiality is critical and essential for for-
warding the cause and benefits for humanitarian logistics. In addition
to these fundamental factors, another factor that also needs to be part
of model development is the concept of fairness.
Finally, as researchers in this area know it is difficult to study situ-
ations in humanitarian logistics. Some of the reasons can be attributed
to lack of time for participation on behalf of personnel from emergency
agencies. This participation is necessary to understand the problem and
gather data. Given a situation, developing analytical models is within
86 Conclusions and Potential for Future Research

reach of the academic research. The difficulties arise in collection of

data, the critical buy-in from the emergency agencies, long periods of
time in doing that and the necessity of interdisciplinary skill set and
researchers to carry out the research. Some of the disasters occur in
developing countries and hence the response and relief work is carried
out in those locations. Access to these countries can pose difficulty.
In this case validating the model using real data, communication of
implications of the solution to the interested agencies and measuring
the effect of such study are critical yet hard to accomplish. Researchers
who have been termed successful in such research studies have been able
to overcome most of these obstacles by virtue of patience, persistence,
and collaboration.
Some of the opportunities for research that will improve efficiency
and effectiveness of the humanitarian logistics are listed below. This is
by no means an exhaustive list but a beginning of a registry of potential
future research topics which is dynamic in nature.
• Prepositioning: Facility location, inventory management for
speed of distribution, developing user-friendly models
• Characteristics of the humanitarian supply chain: visibility
versus speed, complexity of existing supply chains versus sim-
plification for distribution, vulnerability to disruption, pro-
cess cycle time
• Material flow in a humanitarian supply chain: stochastic
demand, managing supply with uncertainty in scale and
scope, complexity due to multiple resources from mixed
sources such as stand-by suppliers, last mile distribution
• People flow in a humanitarian supply chain: evacuation, clus-
ters of responders, local community capacity
• Objective of the humanitarian supply chain: life-saving emer-
gency response versus sustaining long-term developmental
aid
• Performance measures: definition, establishment
• Information management: supply chain network visibility
and transparency of source-to-sink flows, common database
for on-location provisions across humanitarian agencies
• Collaboration: across and within sectors as well as agencies
 6.3 Future Research 87

As students of science and analytics, academics can use their knowledge

and skills to develop tools for improving the efficiency and effectiveness
of humanitarian logistics — and thus improve disaster and humanitar-
ian responses.
 Acknowledgments

First of all I want to express my gratitude to Charles Corbett for the

support and guidance throughout the process. I would like to extend
my heartfelt thanks to Margaret Brandeau for painstakingly reviewing
the very first draft of this article and the anonymous reviewer whose
valuable insights improved this article significantly in the latter stages.
I am also grateful to Moshe Kress and Javier Salmeron for their com-
ments and suggestions, and numerous scholars who shared their work-
ing papers with me. Finally, I would like to thank the publisher for
giving me this opportunity.

88
 References

[1] A. Ak, J. L. Heier, C. L. Wardell III, O. Ergun, and P. Keskinocak, “Improving

the pan American health organization’s vaccine supply chain,” poster presen-
tation at Conference on Humanitarian Logistics Conference, H. Milton Stew-
art School of Industrial and Systems Engineering at Georgia Tech, Atlanta,
United States, 2009.
[2] D. Alexander, “Towards the development of standards in emergency manage-
ment training and education,” Disaster Prevention and Management, vol. 12,
no. 2, pp. 113–123, 2003.
[3] A. Apte, “Fresh produce supply chain: Design and vulnerability,” Unpublished
research, 2008.
[4] A. Apte and G. Ferrer, “Supply chain for vaccine distribution,” Unpublished
research, 2009.
[5] A. Apte and S. Heath, “A plan for evacuation of disabled people in the face
of a disaster,” Unpublished research, 2009.
[6] O. M. Araz, T. Lant, J. Fowler, and M. Jehn, “A pandemic influenza,” poster
presentation at Conference on Humanitarian Logistics Conference, H. Mil-
ton Stewart School of Industrial and Systems Engineering at Georgia Tech,
Atlanta, United States, 2009.
[7] S. Aviles, E. Bah, M. Jimenez, L. Li, A. Morales, J. Wade, and O. Ergun,
“Supply chain optimization for world food programme,” Poster presenta-
tion at Conference on Humanitarian Logistics Conference, H. Milton Stew-
art School of Industrial and Systems Engineering at Georgia Tech, Atlanta,
United States, 2009.

89
90 References

[8] D. L. Bakuli and J. M. Smith, “Resource allocation in state-dependent emer-

gency evacuation networks,” European Journal of Operation Research, vol. 89,
pp. 543–555, 1996.
[9] A. Balakrishanan and J. Geunes, “Collaboration and coordination in
supply chain management and e-commerce,” Production and Operations
Management, vol. 13, no. 1, pp. 1–2, 2004.
[10] B. Balcik and B. M. Beamon, “Facility location in humanitarian relief,”
International Journal of Logistics: Research & Applications, vol. 11, no. 2,
pp. 101–121, 2008.
[11] G. L. Barbarosoglu and Y. Arda, “A two-stage stochastic programming frame-
work for transportation planning in disaster response,” Journal of the Opera-
tional Research Society, vol. 55, pp. 43–53, 2004.
[12] G. L. Barbarosoglu, L. Ozdamar, and A. Cevik, “An interactive approach
for hierarchical analysis of helicopter logistics in disaster relief operations,”
European Journal of Operational Research, vol. 140, pp. 118–133, 2002.
[13] B. M. Beamon, “Measuring supply chain performance,” International Journal
of Operations & Production Management, vol. 19, no. 3, no. 3, pp. 275–292,
1999.
[14] B. M. Beamon, “Humanitarian relief chains: Issues and challenges,” in Pro-
ceedings of the 34th International Conference on Computers & Industrial
Engineering, pp. 77–82, 2004.
[15] B. M. Beamon and S. A. Kotleba, “Inventory management support systems for
emergency humanitarian relief operations in South Sudan,” The International
Journal of Logistics Management, vol. 17, no. 2, pp. 187–212, 2006.
[16] B. M. Beamon and S. A. Kotleba, “Inventory modeling for complex emer-
gencies in humanitarian relief operations,” International Journal of Logistics:
Research and Applications, vol. 9, no. 1, pp. 1–18, 2006b.
[17] T. Boyaci and G. Gallego, “Supply chain coordination in a market with cus-
tomer service competition,” Production and Operations Management, vol. 13,
no. 1, pp. 3–22, 2004.
[18] J. Bramel and D. Simchi-Levi, “On the effectiveness of set partitioning formu-
lations for the vehicle routing problem,” Operations Research, vol. 45, pp. 295–
301, 1997.
[19] M. L. Brandeau and S. S. Chiu, “An overview of representative problems in
location research,” Management Science, vol. 35, no. 6, pp. 645–674, 1989.
[20] M. L. Brandeau, J. H. McCoy, N. A. Hupert, J. E. Holty, and D. M. Bra-
vata, “Recommendations for modeling disaster responses in public health and
medicine: A position paper of the Society for Medical Decision Making,” Med-
ical Decision Making, vol. 29, pp. 438–460, 2009.
[21] D. M. Bravata, G. S. Zaric, J. C. Holty, M. L. Brandeau, E. R. Wilhelm,
K. M. McDonald, and D. K. Owens, “Reducing mortality from anthrax bioter-
rorism: Strategies for stockpiling and dispensing medical and pharmaceutical
supplies,” Biosecurity and Bioterrorism: Biodefense Strategy, Practice, and
Science, vol. 4, no. 3, pp. 244–262, 2006.
[22] G. Brown and W. M. Carlyle, “Optimizing the US Navy’s combat logistics
force,” Naval Research Logistics, vol. 55, pp. 800–810, 2008.
 References 91

[23] J. A. Carbajal, O. Ergun, P. Keskinocak, A. Siddhanthi, and M. Villar-

real, “Debris management operations,” Poster presentation at Conference on
Humanitarian Logistics Conference, H. Milton Stewart School of Industrial
and Systems Engineering at Georgia Tech, Atlanta, United States, 2009.
[24] M. Cataldi, C. Cho, C. Guterriez, J. Hull, P. Kim, A. Park, J. Pickering, and
J. Swann, “Operations research bites back: Improving malaria interventions
in Africa,” Unpublished research, 2009.
[25] Y. C. Chiu and H. Zheng, “Real-time mobilization decision for multi-priority
emergency response resources and evacuation groups: Model formulation and
solution,” Transportation Research: Part E, vol. 43, pp. 710–736, 2007.
[26] R. Church and C. ReVelle, “The maximal covering location problem,” Papers
of the Regional Science Association, vol. 32, pp. 101–118, 1974.
[27] J. A. Cooke, “Are you ready for the next Katrina?,” Logistics Management,
vol. 44, no. 10, p. 80, October 2005.
[28] J. Dai, S. Wang, and X. Yang, “Computerized support systems for emergency
decision making,” Annals of Operations Research, vol. 51, pp. 315–325, 1994.
[29] A. Daoudi, “The united nations world food programme,” Presentation at Con-
ference on Humanitarian Logistics Conference, H. Milton Stewart School of
Industrial and Systems Engineering at Georgia Tech, Atlanta, United States,
2009.
[30] J. M. Day, I. Junglas, and L. Silva, “Information flow impediments in disaster
relief supply chains,” Unpublished research 2009.
[31] B. K. De, D. Herzog, and C. Davila, “Logistics capabilities improvement
project of CARE,” Poster presentation at Conference on Humanitarian Logis-
tics Conference, H. Milton Stewart School of Industrial and Systems Engi-
neering at Georgia Tech, Atlanta, United States, 2009.
[32] V. DeAngelis, M. Mecoli, C. Nikoi, and G. Storchi, “Multiperiod integrated
routing and scheduling of world food programme cargo planes in Angola,”
Computers and Operations Research, vol. 34–36, no. 6, pp. 1601–1615, 2007.
[33] J. Dekle, M. S. Lavieri, E. Martin, H. Emir-Farinas, and R. L. Francis, “A
Florida county locates disaster recovery centers,” Interfaces, vol. 35, no. 2,
pp. 133–139, March–April 2005.
[34] P. J. Denning, “Hastily formed networks,” Unpublished research, April 2006.
[35] Department of the Army, Field Manual 100-16: Army Operational Support.
Washington DC: Headquarters, 1995.
[36] S. Duran, M. Gutierrez, and P. Keskinocak, “Pre-positioning of emergency
items worldwide for CARE international,” Unpublished research, 2008.
[37] P. Eggenhofer, R. K. Huber, B. Katzy, U. Lechner, and S. Richter, “Towards a
research agenda for collaborative crisis response management,” Poster presen-
tation at Conference on Humanitarian Logistics Conference, H. Milton Stew-
art School of Industrial and Systems Engineering at Georgia Tech, Atlanta,
United States, 2009.
[38] D. P. Eisenman, K. M. Cordasco, S. Asch, J. F. Golden, and D. Glik, “Dis-
aster planning and risk communication with vulnerable communities: Lessons
from hurricane Katrina,” American Journal of Public Health, vol. 97, no. S1,
pp. S109–S115, 2007.
92 References

[39] A. Ekici, P. Keskinocak, and J. L. Swann, “Modelling influenza pandemic,

intervention strategies, and food distribution,” under review, Manufacturing
and Service Operations Management, 2010.
[40] B. Eksioglu, M. Jin, I. Capar, Z. Zhang, and S. D. Eksioglu, “Highway traf-
fic management in incidents of national significance,” Journal of Emergency
Management, vol. 6, no. 1, pp. 23–36, January 2008.
[41] B. Emrey, “Health system strengthening for developing countries,” Presenta-
tion at Conference on Humanitarian Logistics Conference, H. Milton Stewart
School of Industrial and Systems Engineering at Georgia Tech, Atlanta, United
States, 2009.
[42] F. G. Engineer, P. Keskinocak, and L. K. Pickering, “Catch-up scheduling for
childhood immunization,” Unpublished research, 2009.
[43] O. Ergun, J. L. Heier, and J. Swann, “Providing information to improve the
performance of decentralized logistics systems,” Working Paper, H. Milton
Stewart School of Industrial and Systems Engineering, Georgia Institute of
Technology, December 2008.
[44] O. Ergun, G. Karakus, P. Keskinocak, J. Swann, and M. Villareal, “Overview
of supply chains for humanitarian logistics,” Unpublished research, 2009.
[45] B. S. Farmer and E. M. Johnson, “NetHope — Collaborating for the future
of relief and development,” Case #6-0026, Tuck School of Business at Dart-
mouth, United States, 2007.
[46] T. Feng and L. R. Keller, “A multiple objective decision analysis for terror-
ism protection: Potassium iodide distribution in nuclear incidents,” Decision
Analysis, vol. 3, no. 2, pp. 76–93, 2006.
[47] C. H. Fine, “Clockspeed-based strategies for supply chain design,” Production
and Operations Management, vol. 9, no. 3, pp. 213–221, 2000.
[48] M. Fisher, J. H. Hammond, W. Obermeyer, and A. Raman, “Configuring
a supply chain to reduce the cost of demand uncertainty,” Production and
Operations Management, vol. 6, no. 3, pp. 211–225, 1997.
[49] J. Fitzsimmons, “A methodology for emergency ambulance deployment,”
Management Science, vol. 19, pp. 627–636, February 1973.
[50] X. Gan, S. P. Sethi, and H. Yan, “Coordination of supply chains with
risk-averse agents,” Production and Operations Management, vol. 13, no. 2,
pp. 135–149, 2004.
[51] X. Gan, S. P. Sethi, and H. Yan, “Channel coordination with a risk-neutral
supplier and a downside risk-adverse retailer,” Production and Operations
Management, vol. 14, no. 1, no. 1, pp. 80–89, 2005.
[52] D. Gazashvili, “CARE International: Emergency preparedness, response, and
recovery,” Presentation at Conference on Humanitarian Logistics Conference,
H. Milton Stewart School of Industrial and Systems Engineering at Georgia
Tech, Atlanta, United States, 2009.
[53] D. E. Gibbons, Communicable Crises. Charlotte, NC: Information Age Pub-
lishing, 2007.
[54] E. Gralla, J. Goentzel, and C. Fine, “Heuristics for emergency response supply
chains,” Poster presentation at Conference on Humanitarian Logistics Con-
ference, H. Milton Stewart School of Industrial and Systems Engineering at
Georgia Tech, Atlanta, United States, 2009.
 References 93

[55] J. Griffin, “Logistics of deworming drug distribution in Zambia,” Poster

presentation at Conference on Humanitarian Logistics Conference, H. Mil-
ton Stewart School of Industrial and Systems Engineering at Georgia Tech,
Atlanta, United States, 2009.
[56] D. Guha-Sapir, D. Hargitt, and P. Hoyois, “Thirty Years of Natural Dis-
asters 1974–2003: The Numbers,” Centre for Research on the Epidemiology
of Disasters, Brussels, http://www.emdat.be/Documents/Publications/ pub-
lication 2004 emdat.pdf. Accessed 5/28/09, 2004.
[57] C. Gunes, W. van Hoevey, and S. Tayur, “An analysis of greater Pittsburgh
community food bank,” Poster presentation at Conference on Humanitarian
Logistics Conference, H. Milton Stewart School of Industrial and Systems
Engineering at Georgia Tech, Atlanta, United States, 2009.
[58] A. Haghani and S. C. Oh, “Formulation and solution of a multi-commodity,
multi-modal network flow model for disaster relief operations,” Transportation
Research, vol. 30, no. 3, pp. 231–250, 1996.
[59] S. Hakimi, “Optimum distribution of switching centers in a communications
network and some related graph theoretic problems,” Operations Research,
vol. 13, p. 462, May–June 1965.
[60] T. Hale and C. R. Moberg, “Improving supply chain disaster preparedness:
A decision process for secure site location,” International Journal of Physical
Distribution & Logistics Management, vol. 35, no. 3, pp. 195–207, 2005.
[61] R. Halper and S. Raghavan, “Efficient utilization of mobile facilities in humani-
tarian logistics,” Poster presentation at Conference on Humanitarian Logistics
Conference, H. Milton Stewart School of Industrial and Systems Engineering
at Georgia Tech, Atlanta, United States, 2009.
[62] C. L. Heidtke, “Reducing the gap of pain: A strategy for optimizing federal
resource availability in response to major incidents,” Master’s Thesis, Naval
Postgraduate School, Monterey, California, 2007.
[63] J. Holguin-Veras and L. Destro, “Estimating material convergence: Flow
of donations for hurricane Katrina,” Poster presentation at Conference on
Humanitarian Logistics Conference, H. Milton Stewart School of Indus-
trial and Systems Engineering at Georgia Tech, Atlanta, United States,
2009.
[64] J. Holguin-Veras, M. Jaller, S. Ukkusuri, M. Brom, C. Torres, T. Wachten-
dorf, and B. Brown, “An analysis of the immediate resource requirements after
hurricane Katrina: Policy implications for disaster response,” Poster presenta-
tion at Conference on Humanitarian Logistics Conference, H. Milton Stewart
School of Industrial and Systems Engineering at Georgia Tech, Atlanta, United
States, 2009.
[65] J. Holguin-Veras, N. Perez, S. Ukkusuri, T. Wachtendorf, and B. Brown,
“Emergency logistics issues impacting the response to Katrina: A synthesis
and preliminary suggestions for improvement,” Poster presentation at Con-
ference on Humanitarian Logistics Conference, H. Milton Stewart School of
Industrial and Systems Engineering at Georgia Tech, Atlanta, United States,
2009.
94 References

[66] N. Hupert, “CDC’s new preparedness modeling initiative: Beyond (and before)
crisis response,” Presentation at Conference on Humanitarian Logistics Con-
ference, H. Milton Stewart School of Industrial and Systems Engineering at
Georgia Tech, Atlanta, United States, 2009.
[67] L. M. Hwee and N. G. Y. Calvin, “An integrated architecture to support
hastily formed network,” Master’s Thesis, Naval Postgraduate School, Mon-
terey, California, 2007.
[68] W. Hyde, “Applied humanitarian logistics,” Presentation at Conference on
Humanitarian Logistics Conference, H. Milton Stewart School of Industrial
and Systems Engineering at Georgia Tech, Atlanta, United States, 2009.
[69] International Committee of Red Cross, http://www.ifrc.org/Docs/pubs/who/
at a glance-en.pdf. Accessed 5/28/09, 2009.
[70] M. Islam, “Country level supply chain process review in NGO operations,”
Poster presentation at Conference on Humanitarian Logistics Conference, H.
Milton Stewart School of Industrial and Systems Engineering at Georgia Tech,
Atlanta, United States, 2009.
[71] A. Iyer, P. McGee, S. Vempala, P. Biswas, and A. Scott, “A web-based system
for homeless shelters,” Poster presentation at Conference on Humanitarian
Logistics Conference, H. Milton Stewart School of Industrial and Systems
Engineering at Georgia Tech, Atlanta, United States, 2009.
[72] M. Jin and B. Eksioglu, “Optimal routing of vehicles with communication
capabilities in disasters,” Unpublished research, 2008.
[73] C. Johnson, “Environmental sustainability at medshare international,” Poster
presentation at Conference on Humanitarian Logistics Conference, H. Mil-
ton Stewart School of Industrial and Systems Engineering at Georgia Tech,
Atlanta, United States, 2009.
[74] S. Khan and A. Richter, “Pilot model: Judging alternate modes of dispensing
in Los Angeles county,” Interfaces, vol. 39, pp. 228–240, 2009.
[75] S. Khan and A. Richter, “Using decision analysis to select alternate modes of
dispensing — An example from Los Angeles county public health,” Journal
of Emergency Management, vol. March/April, pp. 39–51, 2009.
[76] K. King and J. A. Muckstadt, “Public health emergency supply chain models,”
Unpublished Research, 2009.
[77] P. R. Kleindorfer and G. H. Saad, “Managing disruption risks in supply
chains,” Production & Operations Management, vol. 14, no. 1, pp. 53–68, 2005.
[78] P. R. Kleindorfer and L. N. Van Wassenhove, “Risk management for global
supply chains,” in an overview in The Alliance on Globalizing, Chapter 12,
(H. Gatignon and J. Kimberly, eds.), Cambridge University Press, 2004.
[79] A. Klose and A. Drexl, “Facility location models for distribution system
design,” European Journal of Operational Research, vol. 162, no. 1, pp. 4–
29, 2005.
[80] L. Kopczak and E. M. Johnson, “Can heroes be efficient? Information tech-
nology at the International Federation of the Red Cross,” Case #6-0021, Tuck
School of Business at Dartmouth, United States, 2004.
[81] P. Kouvelis and J. Li, “Flexible backup supply and the management of
lead-time uncertainty,” Production & Operations Management, vol. 17, no. 2,
pp. 184–199, 2008.
 References 95

[82] G. Kovacs and K. M. Spens, “Humanitarian logistics in disaster relief oper-

ations,” International Journal of Physical Distribution & Logistics Manage-
ment, vol. 37, no. 2, pp. 99–114, 2007.
[83] M. Kress, Operational Logistics: The Art and Science of Sustaining Military
Operations. Boston: Kluwer Academic Publishers, 2002.
[84] O. Kulemeka, “Using the internet to communicate disaster information to
individuals with disabilities and chronic illness,” Poster presentation at Con-
ference on Humanitarian Logistics Conference, H. Milton Stewart School of
Industrial and Systems Engineering at Georgia Tech, Atlanta, United States,
2009.
[85] H. C. Kunreuther and E. O. Michel-Kerjan, At War With the Weather. Penn-
sylvania: The MIT Press, June 2009.
[86] E. K. Lee, S. Maheshwary, J. Mason, and W. Glisson, “Decision support sys-
tem for mass dispensing of medications for infectious disease outbreaks and
bioterrorist attacks,” Annals of Operations Research, vol. 148, pp. 25–53, 2006.
[87] E. K. Lee, S. Maheshwary, J. Mason, and W. Glisson, “Large-scale dispens-
ing for emergency response to bioterrorism and infectious-disease outbreak,”
Interfaces, vol. 36, no. 6, pp. 591–607, November-December 2006.
[88] E. K. Lee, H. K. Smalley, Y. Zhang, F. Pietz, and B. Benecke, “Facility loca-
tion and multi-modality mass dispensing strategies and emergency response
for biodefense and infectious disease outbreaks,” International Journal on
Risk Assessment and Management — Biosecurity Assurance in a Threaten-
ing World: Challenges, Explorations, and Breakthroughs, vol. 12, no. 2/3/4,
pp. 311–351, 2009.
[89] H. L. Lee, “Aligning supply chain strategies with products uncertainties,”
California Management Review, vol. 44, no. 3, pp. 105–119, 2002.
[90] H. L. Lee, “The triple-A supply chain,” in Harvard Business Review on Supply
Chain Management, pp. 87–115, Harvard Business School Press, 2006.
[91] M. Levi, P. R. Kleindorfer, and D. J. Wu, “Codifiability, relationship-specific
information technology investment, and optimal contracting,” Journal of
Management Information Systems, vol. 20, no. 2, pp. 77–100, 2001.
[92] J. Levins, R. Samii, and L. N. Van Wassenhove, “Fuels: A humanitarian neces-
sity in 2003 post-conflict Iraq; The role of the United Nations joint logistic
center,” No. 07/2005-5290. INSEAD, Fontainebleau, France, 2005.
[93] C. Li and P. Kouvelis, “Flexible and risk-sharing supply contracts under price
uncertainty,” Management Science, vol. 45, no. 10, pp. 1378–1398, 1999.
[94] C. Liang, C. Wang, H. Luh, and P. Hsu, “Disaster avoidance mechanism
for content-delivering service,” Computers & Operations Research, vol. 36,
pp. 27–39, 2009.
[95] E. J. Lodree Jr. and S. Taskin, “Supply chain planning for hurricane response
with wind speed information updates,” Computers & Operations Research,
vol. 36, no. 1, pp. 2–15, 2009.
[96] J. I. MacLellan and D. Martell, “Basing airtankers for forest fire control in
Ontario,” Operations Research, vol. 44, no. 5, pp. 677–686, 1996.
[97] V. Marianov and C. ReVelle, “Siting emergency services,” in Facility Loca-
tion: A Survey of Applications and Methods, (Z. Drezner, ed.), pp. 199–223,
NewYork: Springer-Verlag Inc, 1995.
96 References

[98] A. P. Martinez, C. E. Mejia, O. Stapleton, and L. N. Van Wassenhove,

“Grinding out an effective disaster response in Colombia’s coffee region,” No.
09/2009-5610. INSEAD, Fontainebleau, France, 2009.
[99] A. P. Martinez, O. Stapleton, and L. N. Van Wassenhove, “Last mile fleet man-
agement in humanitarian operations: A case-study based approach,” Unpub-
lished research, 2009.
[100] A. P. Martinez and L. N. Van Wassenhove, “Vehicle replacement in the inter-
national committee of the red cross,” Unpublished research, 2008.
[101] M. Marx, “Coordinating international response to humanitarian crises,” Pre-
sentation at Conference on Humanitarian Logistics Conference, H. Milton
Stewart School of Industrial and Systems Engineering at Georgia Tech,
Atlanta, United States, 2009.
[102] B. Maskell, “The age of agile manufacturing,” Supply Chain Management; An
International Journal, vol. 6, no. 1, pp. 5–11, 2001.
[103] J. Mason and M. Washington, “Optimizing staff allocation in large-scale dis-
pensing centers,” Center for Disease Control and Prevention Report, 2003.
[104] E. L. Maspero and H. W. Ittmann, “The rise of humanitarian logistics,” in
Proceedings of the 27th Southern African Transport Conference, pp. 175–184,
July 2008.
[105] V. M. McCall, “Designing and prepositioning humanitarian assistance pack-up
kits (HA PUKs) to support pacific fleet emergency relief operations,” Master’s
Thesis, Naval Postgraduate School, Monterey, California, 2006.
[106] B. McCorry, “Intra- and inter-organizational collaboration in disaster planning
and long term humanitarian aid,” Presentation at Conference on Humanitar-
ian Logistics Conference, H. Milton Stewart School of Industrial and Systems
Engineering at Georgia Tech, Atlanta, United States, 2009.
[107] J. H. McCoy, “Humanitarian response: Improving logistics to save lives,”
American Journal of Disaster Medicine, vol. 3, no. 5, pp. 283–293, 2008.
[108] J. H. McCoy, “Three central stockpiles for 33 million beneficiaries: UNHCR’s
inventory challenge,” Poster presentation at Conference on Humanitarian
Logistics Conference, H. Milton Stewart School of Industrial and Systems
Engineering at Georgia Tech, Atlanta, United States, 2009.
[109] M. Menezes and M. L. Varela, “Analysis of MSF-Spain stock levels for ongo-
ing projects and the option of creating a new distribution center in east
Africa,” Poster presentation at Conference on Humanitarian Logistics Con-
ference, H. Milton Stewart School of Industrial and Systems Engineering at
Georgia Tech, Atlanta, United States, 2009.
[110] K. Moinzadeh and S. Nahmias, “A continuous review model for an inventory
system with two supply modes,” Management Science, vol. 34, pp. 761–773,
1988.
[111] G. Mortensen, Speech at Naval Postgraduate School. Monterey, California,
12 May, 2009.
[112] J. A. Muckstadt, C. Chan, and J. Chen, “A prototype location allocation
model for emergency response in an Anthrax attack,” Unpublished Research,
2009.
 References 97

[113] V. G. Narayanan and A. Raman, “Aligning incentives in supply chains,” in

Harvard Business Review on Supply Chain Management, pp. 171–193, Harvard
Business School Press, 2006.
[114] P. Nieburg, R. J. Waldman, and D. M. Krumm, “Evacuated populations —
Lessons from foreign refugee crises,” The New England Journal of Medicine,
vol. 353, no. 15, pp. 1547–1549, 2005.
[115] R. Oloruntoba and R. Gray, “Humanitarian aid: An agile supply chain?,”
Supply Chain Management, vol. 11, no. 2, pp. 115–120, 2006.
[116] A. Osuntogun, S. Thomas, J. Pitman, S. Basavaraju, B. Mulenga, and S. Vem-
pala, “V2V: Design of a blood flow system,” Poster presentation at Conference
on Humanitarian Logistics Conference, H. Milton Stewart School of Industrial
and Systems Engineering at Georgia Tech, Atlanta, United States, 2009.
[117] S. H. Owen and M. S. Daskin, “Strategic facility location: A review,” European
Journal of Operational Research, vol. 111, pp. 423–447, 1998.
[118] R. C. Owens Jr., “Collaboration in the USAID/DELIVER PROJECT global
health and SCMS HIV/AIDS supply chains,” presentation at Conference on
Humanitarian Logistics Conference, H. Milton Stewart School of Industrial
and Systems Engineering at Georgia Tech, Atlanta, United States, 2009.
[119] L. Ozdamar, E. Ekinci, and B. Kucukyazici, “Emergency logistics planning
in natural disasters,” Annals of Operations Research, vol. 129, pp. 217–245,
2004.
[120] I. S. Papadakis and W. T. Ziemba, “Derivative effects of the 1999 earth-
quake in Taiwan to U. S. personal computer manufacturers,” in Mitigation and
Financing Seismic Risks, (P. R. Kleindorfer and M. R. Sertel, eds.), Kluwer
Academic Publishers, 2001.
[121] R. Parnell, “American red cross disaster logistics,” Presentation at Conference
on Humanitarian Logistics Conference, H. Milton Stewart School of Industrial
and Systems Engineering at Georgia Tech, Atlanta, United States, 2009.
[122] D. Paton, “Stress in disaster response: A risk management approach,” Disaster
Prevention and Management, vol. 12, no. 3, pp. 203–209, 2003.
[123] H. Peck, “Drivers of supply chain vulnerability: An integrated framework,”
International Journal of Physical Distribution & Logistics Management,
vol. 35, no. 4, pp. 210–232, 2005.
[124] S. J. Pettit and A. K. C. Beresford, “Emergency relief logistic: An evaluation of
military, non-military and composite response models,” International Journal
of Logistics: Research and Applications, vol. 8, no. 4, pp. 313–331, 2005.
[125] A. K. Rathi, R. L. Church, and R. S. Solanki, “Allocating resources to sup-
port a multicommodity flow with time windows,” Logistics and Transportation
Review, vol. 28, no. 2, no. 2, pp. 167–188, 1992.
[126] C. G. Rawls and M. A. Turnquist, “Pre-positioning of emergency supplies
for disaster response,” Paper presented at IEEE International Symposium on
Technology and Society, New York, NY, 2006.
[127] E. Regnier, “Public evacuation decisions and hurricane track uncertainty,”
Management Science, vol. 54, no. 1, pp. 16–28, January 2008.
[128] E. Regnier and P. A. Harr, “A dynamic decision model applied to hurricane
landfall,” Weather and Forecasting, vol. 21, pp. 764–780, October 2006.
98 References

[129] C. ReVelle, D. Bigman, D. Schilling, J. Cohon, and R. Church, “Facility loca-

tion: A review of context-free and EMS models,” Health Services Research,
vol. 12, no. 2, pp. 129–146, 1977.
[130] C. ReVelle, D. Marks, and J. C. Liebman, “An analysis of private and public
sector location models,” Management Science, vol. 16, no. 11, pp. 692–707,
July 1970.
[131] G. H. Saad, “Transformation of supply chain management challenges into
business opportunities,” Unpublished research, 2003.
[132] J. Salmeron and A. Apte, “Stochastic optimization for natural disaster asset
prepositioning,” Unpublished research, 2009.
[133] R. Samii and L. N. Van Wassenhove, “Logistics moving the seeds of a brighter
future (UNJLC’s second year in Afghanistan),” No. 09/2003-5135. INSEAD,
Fontainebleau, France, 2003.
[134] R. Samii and L. N. Van Wassenhove, “The United Nations Joint Logistics
Centre: The Afghanistan crisis,” No. 052003-5092. INSEAD, Fontainebleau,
France, 2003.
[135] R. Samii and L. N. Van Wassenhove, “The United Nations Joint Logistics Cen-
tre (UNJLC): The Genesis of a Humanitarian Relief Coordination Platform,”
No. 04/2003-5093. INSEAD, Fontainebleau, France, 2003.
[136] R. Samii, L. N. Van Wassenhove, K. Kumar, and I. Becerra-Fernandez,
“Choreographer of disaster management: Preparing for tomorrow’s disasters,”
No. 06/2002-5039 INSEAD, Fontainebleau, France, 2002.
[137] R. Samii, L. N. Van Wassenhove, K. Kumar, and I. Becerra-Fernandez,
“Choreographer of disaster management: The Gujarat earthquake,”
No. 602/046/1. INSEAD, Fontainebleau, France, 2002.
[138] F. Sayyady and S. D. Eksioglu, “Optimizing the use of public transit system
during no-notice evacuations of urban areas,” Working Paper, Department of
Industrial & Systems Engineering, Mississippi State University, June 2008.
[139] J. D. Schindler and R. Radichel, “Decision-support for mass vaccination
during a pandemic influenza using agent-based modeling,” Poster presenta-
tion at Conference on Humanitarian Logistics Conference, H. Milton Stewart
School of Industrial and Systems Engineering at Georgia Tech, Atlanta, United
States, 2009.
[140] J. B. Sheu, “An emergency logistics distribution approach for quick response
to urgent relief demand in disasters,” Transportation Research: Part E, vol. 43,
pp. 687–709, 2007.
[141] E. A. Silver, D. F. Pyke, and R. Peterson, Decision Systems for Inventory
Management and Production Planning. New York: Wiley, 3rd ed., 1998.
[142] W. Smith, “Logistics transformation,” Presentation at Conference on Human-
itarian Logistics Conference, H. Milton Stewart School of Industrial and Sys-
tems Engineering at Georgia Tech, Atlanta, United States, 2009.
[143] W. Smith, Private communication. 2009.
[144] S. Sotnikov, “Investigating the effects of partnerships on local health depart-
ments preparedness,” Poster presentation at Conference on Humanitarian
Logistics Conference, H. Milton Stewart School of Industrial and Systems
Engineering at Georgia Tech, Atlanta, United States, 2009.
 References 99

[145] G. C. Souza, Z. Zhao, M. Chen, and M. O. Ball, “Coordinating sales and

raw material discounts in a global supply chain,” Production and Operations
Management, vol. 13, no. 1, pp. 4–45, 2004.
[146] O. Stapleton, A. P. Martinez, and L. N. Van Wassenhove, “Last mile fleet
management in the International Federation of Red Cross and Red Crescent
Societies,” Unpublished research, 2009.
[147] O. Stapleton, A. P. Martinez, and L. N. V. Wassenhove, “Fleet care: Serviciz-
ing in the humanitarian world,” No. 09/2009-5631. INSEAD, Fontainebleau,
France, 2009.
[148] B. Steckler, B. L. Bradford, and S. Urrea, “Hastily formed networks for com-
plex humanitarian disasters: After action report and lessons learned from the
Naval Postgraduate School’s response to hurricane Katrina,” Naval Postgrad-
uate School, Monterey, California, 2005.
[149] D. Stouffer, “Logistics in humanitarian assistance operations,” 2008.
[150] B. C. Tansel, R. L. Francis, and T. J. Lowe, “Location on networks: A survey.
Part I: The p-center and p-median problems,” Management Science, vol. 29,
no. 4, pp. 482–497, April 1983.
[151] E. S. Tean, “Optimized positioning of pre-disaster relief force and assets,”
Master’s Thesis, Naval Postgraduate School, Monterey, California, 2006.
[152] The Economist Global Agenda, “Quality over quantity,” www.economist.com/
PrinterFriendly.cfm?Story ID=3535065. Accessed 5/28/09, 2005.
[153] A. Thomas, “Humanitarian logistics: Enabling disaster response,” Fritz Insti-
tute, pp. 15, 2003.
[154] A. Thomas, “Matching recognition with responsibility,” Freight and Logis-
tics, http://www.fritzinstitute.org/PDFs/InTheNews/2005/ADR 0605.pdf.
Accessed 05/29/09, June 21 2005.
[155] A. Thomas and L. R. Kopezak, “Life-saving supply chains: Challenges and
path forward,” in Building Supply Chain Excellence in Emerging Economies,
(H. L. Lee and C. Y. Lee, eds.), New York: Springer Science, 2007.
[156] A. Thomas and M. Mizushima, “Logistics training: Necessity or luxury?,”
Forced Migration Review, vol. 22, pp. 60–61, January 2005.
[157] B. Thornton, “Disaster preparedness, response, and post-disaster operations,”
Presentation at Conference on Humanitarian Logistics Conference, H. Mil-
ton Stewart School of Industrial and Systems Engineering at Georgia Tech,
Atlanta, United States, 2009.
[158] R. M. Tomasini and L. N. Van Wassenhove, Humanitarian Logistics. New
York: INSEAD Business Press, 2009.
[159] R. M. Tomasini and L. N. V. Wassenhove, “Coordinating disaster logistics
after El Salvador’s earthquake using SUMA’s humanitarian supply manage-
ment system,” Working Paper No. 10/2003-5145. INSEAD, Fonainebleau,
France, 2003.
[160] R. M. Tomasini and L. N. V. Wassenhove, “Genetically modified food dona-
tions and the cost of neutrality: Logistics response to the 2002 food crisis in
Southern Africa,” Working Paper No. 03/2004-5169. INSEAD, Fontainebleau,
France, 2004.
100 References

[161] C. Toregas, R. Swain, C. ReVelle, and L. Bergman, “The location of emergency

service facilities,” Operations Research, vol. 19, no. 6, pp. 1363–1373, October
1971.
[162] P. A. Trunick, “Logistics when it counts,” Logistics Today, vol. 46, no. 2, p. 38,
2005.
[163] E. Uzun, N. T. Argon, and S. Ziya, “A mathematical approach to triage in the
context of emergency response planning,” Poster presentation at Conference
on Humanitarian Logistics Conference, H. Milton Stewart School of Industrial
and Systems Engineering at Georgia Tech, Atlanta, United States, 2009.
[164] L. N. Van Wassenhove, “Humanitarian aid logistics: Supply chain management
in high gear,” Journal of Operational Research Society, vol. 57, no. 5, pp. 475–
489, 2006.
[165] M. H. Whitworth, “Designing the response to an anthrax attack,” Interfaces,
vol. 36, no. 6, pp. 562–568, November/December 2006.
[166] P. Yadav, “Intra- and inter-organizational collaboration in disaster planning
and long term humanitarian aid: Perspectives from the architecture for financ-
ing global health,” Presentation at Conference on Humanitarian Logistics
Conference, H. Milton Stewart School of Industrial and Systems Engineer-
ing at Georgia Tech, Atlanta, United States, 2009.
[167] W. Yi and L. Ozdamar, “A dynamic logistics coordination model for evacua-
tion and support in disaster response activities,” European Journal of Opera-
tional Research, vol. 179, pp. 1177–1193, 2007.
[168] W. Yushimito, M. Jaller, and S. Ukkusuri, “Application of voronoi based
heuristic for facility location in disasters,” Poster presentation at Conference
on Humanitarian Logistics Conference, H. Milton Stewart School of Industrial
and Systems Engineering at Georgia Tech, Atlanta, United States, 2009.