Research Problem

The problem forming the basis of this research is the process of transporting severely
injured patients to healthcare facilities in a region with numerous casualties following a major
disaster such as an earthquake. The uncertainty of many parameters during this process
significantly complicates the problem, making it more challenging to solve.

This research seeks to answer the following questions:

▪ Can the process of transporting casualties to healthcare facilities in a region with a large
number of injuries after an earthquake be improved?
▪ Can the transport times of casualties be reduced?
▪ Can the total transport time be optimized?

Problem Representation

The details of the problem, the steps of the proposed solution to achieve the desired objectives,
and their characteristics are outlined below:

• The casualty logistics problem addressed in this research is designed as an integrated

logistics network.

• The problem consists of a three-layer integrated logistics network:

o Layer 1: Ambulance stations located at hospitals.
o Layer 2: Severely injured individuals in earthquake-stricken areas.
o Layer 3: Hospitals.

• It is assumed that search and rescue operations will continue for 72 hours.
 • Pendik District of Istanbul Province has been chosen as the application area.

• The locations of casualties within the Pendik District boundaries are selected randomly.

• Hospitals and ambulances located in the Pendik District are utilized.

• In directing casualties to hospitals, hospital occupancy rates varying across periods have
been taken into account.

• The Pendik District of Istanbul has been assessed independently, excluding patients,
ambulances, and hospitals from neighboring districts.

• Incoming patient information is prioritized based on the severity of the injury (triage).
Only severely injured individuals are assigned ambulances.

• The total number of severely injured individuals is assumed to arrive at the system as an
exponential function every hour.

• Shortest duration routes are selected for ambulances to reach patients.

• Once patients are reached, they are transported to hospitals via the shortest route
available.

Problem Parameters

In this research, the following considerations are assessed for the problem:

• Whether it is possible to reach casualties at every location where assistance is requested.

• The time taken to respond to the requests.

• The proportion of requests that are fulfilled.

• The order in which patients are attended to.

• Which patient is transported to which hospital.

• The sufficiency of ambulance capacities to meet the anticipated demand.

• The routes ambulances will take to reach patients.

• The routes ambulances will take to transport patients to hospitals.

• The sufficiency of hospital capacities to meet potential demands.

This ensures the following outcomes:

• Optimum utilization of available hospitals, ambulances, and equipment during disasters.

• Reaching patients in the shortest time possible.

• Transporting patients to hospitals in the shortest possible time.

• Minimizing loss of life.

• Minimizing service delivery times.

• Maximizing the number of patients served.

Characteristics of the problem parameters:

Hospitals:

• Hospital addresses.

• Coordinates.

• Bed capacities.

• Number of medical staff.

• Occupancy rates (historical data).

• Number of inpatients (by year).

Ambulances:

• Locations of ambulances in the Pendik District (coordinates).

• Total number of ambulances.

• Types/classes of ambulances.
Patients:

• Number of patients.

• Locations of patients.

• Triage information.

Assumptions of the Problem

Although ambulances initially depart from their stations, it is observed that they are
unlikely to return to the stations due to the dynamic nature of the process and the high demand.
Ambulances are expected to head directly to the next assigned patient after delivering the
current patient to a hospital. Ambulances will not return to their stations until all patients are
transported. For this reason, all publicly owned ambulances in the Pendik District are assumed
to start their operations from Pendik State Hospital.

The primary assumptions are as follows:

• The critical 72-hour period after an earthquake, during which the probability of survival
is highest, is divided into 72 one-hour periods.

• Considering that individuals rescued from earthquakes are generally classified into three
categories: Minor Injuries, Moderate Injuries, and Severe Injuries.

• In this study, priority for transport to hospitals is given to severely injured individuals.

• The optimization is entirely based on the transport of severely injured individuals.

• It is assumed that minor and moderate injuries receive first aid at the scene without
occupying ambulances or being transported to hospitals.

• Patients who cannot be served in a given period are prioritized for transport in
subsequent periods through penalty costs.

• Ambulances initially transport patients to assigned hospitals until bed capacities are full,
after which they transport patients to other hospitals.

• Severe injury demands for each period are determined using an exponential function,
with high numbers initially decreasing over time.
 • Ambulances unable to deliver patients to assigned hospitals due to capacity issues are
redirected to other hospitals in subsequent periods.

Objective of the Research

One of the most frequent challenges crisis managers face during disasters and
emergencies is the transportation of casualties to hospitals amidst high demand. Determining
the routes to follow for delivering assistance to needed areas, identifying distribution locations,
and accounting for all risk factors along the routes are among the key topics in logistics
network design.

The objective of this research is to design a logistics network that minimizes the time
taken for severely injured patients rescued from earthquakes to be transported to hospitals by
ambulances. A mathematical model has been developed for this purpose, along with a novel
patient selection strategy.

The proposed mathematical solution model aims to optimize the process of transporting
severely injured individuals from earthquake-stricken areas to permanent and temporary
hospitals. The patient selection strategy aims to minimize patients' waiting times and coordinate
transportation operations.

By utilizing the optimization model and the prioritized patient selection strategy, the
goal is to ensure the fastest possible transport of patients to healthcare facilities and reduce the
number of patients waiting.

Constraints of the Research

Key constraints of this study include:

▪ The system is designed solely for land ambulances, without integrating air or sea vehicles,
which constitutes a significant limitation of the study.
▪ Since the application is limited to Pendik District, the idle capacities of neighboring districts
cannot be utilized, marking a critical constraint.
▪ Including nearby hospitals in the system could make the objective function more optimal.
▪ Not evaluating injury demands across the entire city of Istanbul and excluding all hospitals
and ambulances dynamically are significant limitations.
Research Model and Methodology

In examining logistics network design and time-distributed dynamic allocation problems, it is

evident that various mathematical methods are employed based on the type of logistics network
and the structure of the available data.

• The casualty logistics problem addressed in this research was designed as a multi-stage
integrated logistics network. The problem was solved using the proposed MILP model
and patient selection model.

• Dijkstra’s Shortest Path Algorithm was used to determine the shortest distances between
hospitals and casualties.

• The allocation of ambulances to specific patients was determined using MILP.

• The Branch and Cut Algorithm (CBC) was used to solve the model, aiming to minimize
the total transportation time of ambulances.

• The patient selection method determined which patients to attend to and which could
not be reached.
Grafic Model
4.2.3 Notation
Sets and Parameters

For solving post-disaster casualty logistics problems, the necessary sets and parameters are
defined as follows:

Sets and Parameters:

• H : Set of hospitals h=1,2,…..,H

• Hloc : Hospital locations Hloc ∈ H

• Hcap : Hospital capacities Hcap ∈ H

• P : Set of severely injured patients p=1,2……P

• Ploc : Patient locations Ploc ∈ P

• A : Set of ambulances a=1,2……A

• a : Ambulance location a∈H

• t : Time period t=1,2,… .. ,T

• ct : Cost (e.g., waiting cost) ct=1,2…..C

State Variables

• Anum : Number of ambulances operating in period t

• Acap : Ambulance patient transport capacities

• Annp : Patients that ambulances need to transport during a period

• Anap : Patients assigned to ambulances for transportation in a period

• Ansp : Patients transported by ambulances in a period

• Antp : Total patients transported by each ambulance

• Atst : Total service time for each ambulance

 • Atsp : Ambulance service times during a period

• Hcap : Hospital empty capacities at the start of period ttt

• Htcap : Total empty capacities of hospitals at the start of period t

• Hnsp : Number of hospitalized patients with beds at each hospital

• Hnst : Number of patients treated as outpatients at each hospital

• Ptn : Newly injured patients at the start of period t

• Pus : Patients not served in t−1, carried over to period t

• Ptt : Total patients waiting for ambulances at the start of period t

• Ptns : Number of patients served at the end of period t

• Ptnu : Number of patients not served at the end of period t

• Ptwt : Waiting time for each patient to receive service

• Ptst : Total transport time for each patient

• Ta2p : Time for an ambulance to reach a patient

• Ta2h : Time for an ambulance to return to the hospital

• Tatst : Total service time of ambulances

Decision Variable

M :Indicator of whether a patient is assigned to an ambulance

Objective Function

Minimize

𝑧 = ∑𝑇𝑡=1 ∑𝑎∈𝐻 ∑𝑝∈𝐻 ∑ℎ∈𝐻 𝑐𝑡 (𝑇𝑎2𝑝 + 𝑇𝑎2ℎ )𝑀

Constraints

𝐻
∑ 𝑀 = 𝐴𝑛𝑛𝑝
ℎ=1

∀A∈P

M ∈ {0,1}
M ≥ 0 Patient Assignment Situation

Annp≥ 0 The number of patients that ambulances need to transport in a

period must be greater than or equal to 0.

Ansp≥ 0 The number of patients transported by ambulances in a period

must be greater than or equal to 0.

Annp≥ Ansp The number of patients that need to be transported must be

greater than or equal to the number of transported patients.

Antp≥ 0 The total number of patients transported by each ambulance

must be greater than or equal to 0.

Anum≥ 1 The total number of ambulances must be greater than 1.

Acap=1 The capacity of ambulances is limited to 1 patient.

Ptt≥ 0 The total number of patients waiting at the beginning of period

t must be greater than or equal to 0.

Ptns≥ 0 The number of patients served by the end of period t must be

greater than or equal to 0.
Ptt≥ Ptns The total number of patients waiting at the beginning of period
t must be greater than or equal to the number of patients served
by the end of the period.

Ptwt≥ 0 The waiting time for each patient to receive service must be
greater than 0.

Ptn≥0 The number of new patients at the beginning of period t must

be greater than or equal to 0

Pus≥0 The number of patients left untreated in period t-1 must be

greater than or equal to 0.

Ptt≥ Ptn + Pus The total number of patients waiting at the beginning of period
t is equal to the sum of new patients at the beginning of period
t and the number of untreated patients from period t-1.

Ptt≥ Ptns + Ptnu The total number of patients waiting at the beginning of period
t is equal to the sum of patients transported and patients left
untreated by the end of period t.

Ptt= Annp The total number of patients waiting at the beginning of period
t is equal to the number of patients that ambulances need to
transport during the period.

Hnsp≥0 The number of inpatients treated by each hospital must be

greater than or equal to 0.

Hnst≥0 The number of outpatients treated by each hospital must be

greater than or equal to 0.
 Tatst >0 The total service time of ambulances must be greater than 0.
Ta2p >0 The time for the ambulance to reach the patient must be greater
than 0.
t >1 The number of periods must be greater than 1.
ct >1 The penalty score must be greater than or equal to 1.

Assumptions

• Ambulances can make multiple trips within a single period.

• Ambulances cannot transport patients to hospitals other than the one they are assigned
to during the period.
• Ambulances are stationed at hospitals.
• The critical 72-hour period following the earthquake is divided into hourly periods,
totaling 72 periods.
• Patients are categorized into three groups: Mildly Injured, Moderately Injured, and
Severely Injured.
• The transportation problem is addressed for patients in the severely injured category.
• First aid is provided on-site for mildly and moderately injured individuals without
occupying hospitals or ambulances.