Distribution Planning and Control:

An Experimental Comparison of DRP and Order Point

Replenishment Strategies



S. T. Enns
Pattita Suwanruji




Dept. of Mechanical and Manufacturing Engineering
University of Calgary
Calgary, AB, Canada, T2N 1N4


Abstract:

The increased globalization of trade has resulted in greater emphasis on Supply Chain
Management. The planning and control strategy used to regulate the flow of goods is important
in determining performance. Objectives are often stated in terms of meeting delivery
requirements while minimizing distribution chain inventory. Distribution Requirements
Planning (DRP) and Order Point replenishment systems are two commonly used strategies with
very different information requirements. This research examines the performance within
networks involving manufacturing, distribution and retail facilities. Multiple products with non-
stationary demand are considered. Simulation results are used in identifying strengths and
performance characteristics of each strategy.

Key Words:

Supply Chain Management, Inventory Replenishment, Logistics, Distribution
Requirements Planning (DRP).

-1-
Introduction


Supply Chain Management (SCM) issues have recently been of intense interest to both
researchers and industrial practitioners. Benefits of better management include improved delivery
performance, reduced inventory, and increased flexibility and responsiveness to customer demand
(see 1997 Benchmarking Study, 1997). Handfield and Nichols (1999) give the definition of Supply
Chain Management as the management of information systems, sourcing and procurement,
production scheduling, order processing, inventory management, warehousing, custom service, and
after-market disposition of packaging and materials. Obviously, a major part of Supply Chain
Management is the integration of information from business organizations whose activities are
related to the provision of raw materials, production, and distribution of final products. This paper
presents a preliminary investigation of integrated performance within a stochastic demand
environment. Distribution Requirements Planning (DRP) and Order Point planning and control
systems are compared.
Distribution Requirements Planning (DRP) is a time-phased replenishment approach in
which inventory status is reviewed and new shipment plans are generated periodically. The concepts
and logic used are similar to those used in Material Requirements Planning (MRP). However, the
approach is applied toward the distribution of goods separated geographically rather than the flow of
parts within one manufacturing facility. The demand at the stage most closely linked to the end user
is treated as independent demand, while demand at upstream stages is treated as dependent demand.
Forecasting is used to anticipate requirements at the independent demand level (or echelon), while
time-phased logic is use to anticipate requirements for upstream inventory echelons. Inventory status
and planned lead times are therefore used to time order releases so inventory is minimized while still
preventing excessive shortages. Since inventory is planned throughout the supply chain, based on
anticipated demand, DRP systems are considered to be proactive. Advantages of DRP include
inventory reduction, better customer service, and compatibility with other systems within the supply
chain (e.g. MRP, transportation planning, etc.). Martin (1983), How DRP Helps (1984), Maskell
(1988), Ross (1988), or Ho (1990) provide further information. However, DRP also has
disadvantages. Data throughout the supply chain must be accessible (Weiler, 1998), implementation

-2-
costs are relatively high, and system nervousness can result in highly uncertain environments (Ho,
1992).
Order Point systems, on the other hand, treat demand at all echelons in the supply chain as
independent demand. Control is decentralized and inventory is reviewed at each stage
independently. Under continuous review, orders are placed when the inventory falls below the
reorder point. If inventory falls to zero, backorders are also taken into account in placing new orders.
Reorder points are calculated on the basis of expected demand during expected replenishment times
(actual lead times). Order Point systems are considered to be reactive, since they often use average
information for replenishment decision and do not have mechanisms to anticipate the changes in
demand. This tends to lead to poor performance in environments where demand fluctuation is high.
DRP has been proven to work well where supply chain integration is feasible, especially
when compared to traditional Order Point systems (Cooke (1983), Smith (1985), and Madia (1990)).
However, there has been little research that compares DRP performance to Order Point performance
under similar conditions. In this research, the two planning and control systems are modelled and
compared experimentally, using discrete-event simulation. The objective is to gain further insights
into the behaviour, performance and applicability of these systems.
In this research, delivery performance and inventory requirements are examined with respect
to lot sizes, planned lead times (in DRP systems) and reorder points (in Order Point systems). The
hypothesis is that DRP will perform better than Order Point systems in a supply chain network with
stochastic, uncertain demand. The reason is that DRP is proactive and uses forecast information to
anticpate time-phased requirements throughout the supply chain.
The remainder of the paper is organized as follows. The next section describes the supply
chain network and general methodology used in this research. Following this, the two planning and
control systems are described. Next, the experimental design is presented. Results and analysis
follow this. Finally, the conclusions are stated and future research opportunities are identified.





-3-
The Supply Chain Network

The supply chain modelled consisted of four echelons configured in the network shown in
Figure 1. This network involved eight nodes from which inventory could be shipped or received;
two Suppliers, two Manufacturers, two Distributors and two Retailers. Material flowing between
each of the nodes was designated by a unique part number, shown as Parts 1 through 8 in the
diagram. Only one part type was held at each of the nodes, except at the Retail nodes. These each
stocked two part types. Raw material at the Supplier nodes was assumed to be infinite so that no
shortages or back orders occurred at this inventory echelon.
A discrete-event simulation model of this network was developed using ARENA 3.0 (1994).
This supply chain simulator was interfaced with an Excel spreadsheet program through the use of
Visual Basic for Applications (VBA). The Excel spreadsheet was used to specify all model and
experimental control inputs, including whether DRP or the Order Point system was being used. As
well, the Excel spreadsheet contained extensive Visual Basic macros for creating order release
schedules under Distribution Requirements Planning (DRP). These plans were communicated to the
supply chain simulator on a weekly basis during experimental runs.
Customer orders for P1 through P4 were assumed to be accumulated during each week and
shipped at the end of the week. Forecasts of demand were also made on the basis of weekly
intervals. The forecast weekly demand patterns at the Retail level are illustrated in Figure 2. The
average forecasts values were set to 2000, 1500, 2000 and 2500 units per week for P1 through P4,
respectively. The forecast patterns were sinusiodal with a cycle length of 52 weeks and amplitudes
of 360, 360, 180 and 180 units. The forecast patterns were also offset so that peak demand resulted
at 13, 39, 26 and 0 weeks for Parts 1 through 4, respectively. The actual weekly demands were
normally distributed around the forecast. The coefficient of variation for actual weekly demand
relative to the forecast weekly demand was 0.1.

-4-

Fig. 1 - Supply Chain Network
P1
P2
Customers Demand
P5
P7
Infinite Raw Material for P7
Lots in
Transit
for P7
Orders
Retailer 1
Distributor 1
Manufacturer 1
Supplier 1
P4
P3
Customers Demand
P6
P8
Infinite Raw Material for P8
Lots in
Transit
for P8
Orders
Retailer 2
Distributor 2
Manufacturer 2
Supplier 2
P5 P6
P5 P6
P7 P8
Raw Material
for P7
Raw Material
for P8
P1 = 2000
units/40- hr.
week
P2 = 1500
units/40-hr.
week
P3 = 2000
units/40-hr.
week
P4 = 1500
units/40-hr.
week
Lot Sizes
K = 200, LS = 1265
K = 600, LS = 2191
K = 1000, LS = 2829
Transit Time
Mean = 80 hrs
StdDev = 8 hr
Lot Sizes
K = 200, LS = 1265
K = 600, LS = 2191
K = 1000, LS = 2829
Transit Time
Mean = 80 hrs
StdDev = 8 hr
Lot Sizes
K = 200, LS = 1265
K = 600, LS = 2191
K = 1000, LS = 2829
Transit Time
Mean = 80 hrs
StdDev = 8 hr
Lot Sizes
K = 200, LS = 1265
K = 600, LS = 2191
K = 1000, LS = 2829
Transit Time
Mean = 80 hrs
StdDev = 8 hr
Lot Sizes
K = 200, LS = 895
K = 600, LS = 1550
K = 1000, LS = 2000
Transit Time
Mean = 80 hrs
StdDev = 8 hr
Lot Sizes
K = 200, LS = 775
K = 600, LS = 1342
K = 1000, LS = 1733
Transit Time
Mean = 80 hrs
StdDev = 8 hr
Lot Sizes
K = 200, LS = 895
K = 600, LS = 1550
K = 1000, LS = 2000
Transit Time
Mean = 80 hrs
StdDev = 8 hr
Lot Sizes
K = 200, LS = 1000
K = 600, LS = 1733
K = 1000, LS = 2237
Transit Time
Mean = 80 hrs
StdDev = 8 hr
Lots in
Transit
for P5
Orders
Lots in
Transit
for P6
Orders
Lots in
Transit
for P1
Orders
Lots in
Transit
for P2
Orders
Lots in
Transit
for P3
Orders
Lots in
Transit
for P4
Orders

-5-
The mean, or expected, transit times between all nodes was 80 hours. This was equivalent to
2 weeks under the time unit assumptions made in this research. Actual transit times were normally
distributed, with a coefficient of variation of 0.10. These transit times applied to the shipment of lots
of material, with no lot splitting allowed. The nodes in Figure 1 were assumed to be unconstrained
with respect to capacity, except for the Manufacturing nodes. The two manufacturing facilities were
each modelled as a single resource. The time spent going through each resource was assumed to
represent the aggregate of all operation flowtimes. The utilization of the manufacturing facilities
was in the range of 90% and the flowtimes were approximately 24 hours. Therefore, a
manufacturing time allowance of 24 hours was added to the replenishment time for P5 and P6. In
other words, the expected replenishment time at the Distributor nodes was based on an expected
transit time of 80 hours plus an expected manufacturing time of 24 hours.














Fig. 2 Forecast and Demand Patterns

0
500
1000
1500
2000
2500
3000
3500
52 56 60 64 68 72 76 80 84 88 92 96 100 104
Weeks
W
e
e
k
l
y

D
e
m
a
n
d

a
n
d

F
o
r
e
c
a
s
t
P1-Demand
P1-Forecast
P2-Demand
P2-Forecast
P3-Demand
P3-Forecast
P4-Demand
P4-Forecast

-6-
Planning and Control Systems

The DRP system requires planned lead times and lot sizes to be set, while the Order Point
system requires the reorder point and lot sizes to be set. The inventory-delivery performance tradeoff
of interest in this type of research can be evaluated by adjusting these system decision parameters.
The Fixed Order Quantity (FOQ) lot-sizing rule was used since it is applicable to each of the
planning and control systems. It is desirable to use the same lot-sizing policy, as well as actual lot
sizes, since this eliminates lot sizing as an experimental factor and facilitates making comparisons.
Therefore, performance is controlled on the basis of adjusting planned lead times for DRP and
adjusting the reorder point for Order Point systems.
It should be noted that DRP can be run with a variety of other lot-sizing policies and it is not
being claimed that FOQ is best. Bookbinder and Heath (1988) analyze and compare several other
lot-sizing policies under DRP assumptions. The logic used in FOQ lot-size setting is explained next.
Following this, the order release logic for each of the systems is described.

Calculation of Lot Sizes (Q):

Economic Order Quantity (EOQ) assumptions were used in determining appropriate lot sizes.
Relevant costs include the fixed cost of placing a new shipment order and the holding costs for
inventory stocked at the nodes in the supply chain network. Holding costs for inventory in transit are
not relevant since the expected amount of inventory in transit is constant. As well, transportation
costs are assumed to be a linear function of the shipment size. Therefore, total transportation costs
are also constant, regardless of shipment lot sizes. The following EOQ relationship then applies.
where:
Q
i
*
- optimal lot size for Part i based on EOQ assumptions.
S
i
- order cost, or setup cost, for shipments of Part i.
D
i
- average demand per time unit for Part i
h - inventory holding cost rate per unit time.
v
i
- value, or cost, of part i.

v
h
)
D S
(2
= Q
i
i i *
i

-7-

Now if the order cost is assumed to be equivalent to the cost of holding some amount of
inventory for a given amount of time, then we can set S
i
equal to the following:



Therefore,

As an example, assume the demand for Part i is 2000 units per week and k is set at 1000 unit-weeks.
In this case 2000 = 2000 ) (1000 2 = Q
*
i
.

Logic for DRP order release:

The Distribution Requirements Planning logic was based on a time-bucketed system.
Forecasts for independent demand parts were based on expected demand during intervals of one
week. The planning horizon was set at 10 weeks and unbiased forecasts were generated for each
week in the planning horizon. The distribution plan was regenerated weekly and forecasts were
updated on the basis of a rolling horizon.
The planning bucket for determining order releases was set equal to 0.05 weeks. Therefore,
order releases for any Part i could occur at any one of 20 different times during a week. Many
examples of DRP logic use a planning bucket equal to the plan regeneration interval. However, the
DRP planning bucket can be much smaller than the replanning interval, just as in MRP systems. The
smaller time bucket improves performance as well as making it possible to develop much better
inventory-delivery performance tradeoff curves.
The logic used in DRP is similar to that used in Materials Requirements Planning (MRP).
Forecasts are made at the echelon furthest downstream (or closest to customer demand). This is
often the retail echelon. Orders are planned at this level in a similar manner to how order releases for
independent demand items are planned in an MRP system. Assuming the plan is being generated at
the current time, tp, the following relationship is used to determine the schedule of order releases, Q
i,t
, at discrete future times, T.
)
v
(h k =
S i i
)
D
k (2 = Q
i
*
i

-8-

where:
Q
i ,T
- order quantity scheduled for release at time T.
Q
i
*
- optimal lot size, based on EOQ assumptions.
Forecast
i,t
- forecast of weekly demand for Part i at time t.
SR
i,t
- scheduled receipts at time t of Part i released orders.
PR
i,t
- planned receipts at time t of Part i planned orders.
Inv
i,tp
- quantity of Part i, WIP or finished goods, in stock at
time tp.
BO
i,tp
- quantity of Part i, WIP or finished goods, back
ordered at time tp.
PLT
i
- planned lead time, equal to a multiple of the size of
the planning bucket, for replenishment of Part i.
T - time at the beginning of a discrete planning bucket.
tp - current time, when the plan is generated.

The quantity in square brackets is rounded up to the nearest integer, as indicated by IntRndUp.
The order releases for upstream echelons are derived from gross requirements, based on the
order releases from the downstream echelon which they supply. Orders are planned in a similar
manner to planning dependent demand items in an MRP system. However, the process is sometimes
referred to as one of implosion, rather than explosion. Since there is no Bill of Materials specifying
assembly requirements, as in MRP, the planning process is simply one of moving upstream through
each inventory echelon sequentially. All downstream requirements are aggregated and then the
quantity and timing of orders are calculated using the following relationship.

]
]
]

+ - - -
Q
Q
BO Inv PR SR Forecast
IntRndUp 0, = Q
*
i
*
i
tp i, tp i, t i,
PLT + T
tp = t
t i,
PLT + T
tp = t
t i,
PLT + T
tp = t
T i,
i i i
max

-9-

where:
GR
i,t
- gross requirements for Part i at time t.

The gross requirements for Part i at time t are equal to the sum of all orders for this part at
downstream echelons.
The planned lead times for Part i, PLT
i
, were set equal to the expected replenishment time,
rounded up to the next planning time bucket, multiplied by a safety factor.




where:
PLT
t
- planned lead time for part i
RT
t
- expected time units for replenishment of part i
SF
t
- safety factor for part i
p - length of time in planning bucket

The safety factor, SF
i
, was used to control the tradeoff between the inventory levels and the
delivery performance observed. Increasing SF
i
increases inventory holding costs but reduces tardy
deliveries. It should be noted that DRP systems also lend themselves to the use of safety stock, as
opposed to using this approach of inflating planned lead times. However, safety stock was not
considered in this study since this would introduce an extra experimental control factor applicable to
DRP alone.

Logic for Order Point order release:

The inventory position in the Order Point system was reviewed on a continuous basis. Any
event affecting the system inventory could trigger a new order release. At each change in the
]
]
]

+ - - -
Q
Q
BO Inv PR SR GR
IntRndUp 0, = Q
*
i
*
i
tp i, tp i, t i,
PLT + T
tp = t
t i,
PLT + T
tp = t
t i,
PLT + T
tp = t
T i,
i i i
max
p
SF

p
RT
IntRndUp =
PLT i
i
i

-10-
inventory position, the following relationship was evaluated to determine whether a new order
should be placed and how large this order should be.


where:
Inv
i,t
- quantity of Part i, WIP or finished goods, in stock at time t.
OP
i
- reorder point for Part i
OR
i,t
- quantity of Part i in orders released to the supplier but not
filled, at time t, due to lack of upstream inventory.
OT
i,t
- quantity of Part i in transit to the destination, at time t, where
they are required.
BO
i,t
- quantity of Part i, WIP or finished goods, back ordered at the
destination at time t.
Q
i
*
- optimal lot size determined on the basis of EOQ assumptions.

The order quantity must always be 0, Q
i
*
or a multiple of Q
i
*
.
The reorder point, OP
i
, was set equal to the expected part demand during the expected
replenishment time, multiplied by a safety factor. This relationship can be stated as follows:



where:
D
i
- average demand per time unit for Part i
RT
i
- expected time units for replenishment of Part i
SF
i
- safety factor for Part i

The expected replenishment time, RT
i
, was set equal to the average transit time for shipments
placed between non-manufacturing destinations. If the order was placed to a manufacturing facility,
the expected production time was also taken into account. The safety factors, SF
i
, are used to adjust
the reorder point and vary what is commonly referred to as the safety stock level.


|
|

\
|
]
]
]
]

Q *
Q
)
BO
-
OT
+
OR
+
Inv
( -
OP
IntRndUp 0, = Q
*
i *
i
t i, t i, t i, t i, i
t i,
max
SF
)
RT
(
D
=
OP i i i i

-11-
Experimental Design

The experimental design consisted of three factors. The first factor was the distribution
planning and control system. The settings for this factor were DRP and Order Point. The second
factor was the lot-sizing parameter k. This parameter controls the relationship between inventory
holding costs and order or setup costs. The higher the value of k is, the more important order costs
become relative to inventory holding costs. This affects the optimal lot sizes, Q
i
*
, calculated. Higher
values of k result in larger lot sizes. The values of k used in this research were 200, 600 and 1000
unit-weeks. The resulting lot sizes with k = 200 were 895, 775, 895, 1000, 1265, 1265, 1265 and
1265 for Parts 1 through 8, respectively. For k = 600 they were 1550, 1342, 1550, 1733, 2191,
2191, 2191 and 2191, while for k = 1000 they were 2000, 1733, 2000, 2237, 2829, 2829, 2829 and
2829.
The final experimental factor was the safety factor, SF
i
,used to control delivery performance.
This factor was run at 20 levels for each combination of other factor settings so that inventory-
delivery performance curves could be generated. The safety factor for the DRP system adjusted
planned lead times. It was set at increments of 0.05, between 1.0 and 1.95, and was applied equally
to all parts. The safety factor for the Order Point system was used to adjust the reorder point. In
other words, the safety stock level was being adjusted. This value was also adjusted by increments
of 0.05, between 1.0 and1.95.
This experimental design resulted in 120 simulation settings. One replication was run at each
setting for these preliminary results. However, tests indicated there was little variance in results
when multiple replications were run. Each replication was five years in length, with data from the
first year being discarded to avoid initialization effects.
The primary measures in this research were delivery mean tardiness and inventory part
counts. Tardiness is defined to equal zero if a part is delivered early or on time, and to equal the
lateness otherwise. Delivery tardiness was measured on the basis of deliveries to end-user customers
only, not deliveries at upstream echelons. Therefore, only the delivery performance of Parts 1
through 4 were considered relevant. Inventory part counts were based on the total quantity of parts
in the distribution system. This did not include Supplier inventory, assumed to be unlimited with
respect to replenishment availability, but did include parts in transit from the Supplier echelon.

-12-
Results and Analysis

The customer delivery performance results, based on mean tardiness, versus total system
inventory are summarized in Figures 3 to 7. The points along each line, moving toward the right,
represent increasing safety factor, SF
i
, settings. Therefore, these curves indicate increasing planned
lead times for DRP and increasing reorder points for Order Point systems.
Figures 3 and 4 show the effect of increasing lot sizes for DRP and Order Point systems
respectively. Lot sizes, Q
i
*
, increase as the value of k, which indicates the tradeoff between order
costs and holding costs, increases. In all cases, increasing lot sizes shifts the performance curves to
the right. This is consistent with expectations. As order costs increase relative to holding costs,
larger lot sizes are justified. This means delivery performance will be worse for a given inventory
level or, alternately, inventory will be higher for a given level of delivery performance.
Of more interest is the relative performance of DRP and Order Point systems. These results
are compared in Figures 5 through 7. The graphs show that DRP consistently performs better than
the Order Point system. These results support the hypothesis that centralized planning and control,
based on global inventory and order information, performs best when demands vary through time
and when there is significant uncertainty with respect to demand and replenishment times. This
superior performance is believed to be due to the ability of DRP to anticipate changes, based on
forecast information, in demand along the supply chain and release time-phased orders in
anticipation of future requirements.




-13-












Fig. 3 - Planned Lead Time and Lot Sizes Effect for DRP















Fig. 4 - Reorder Point and Lot Sizes Effect for Order Point









0.0
2.0
4.0
6.0
8.0
10.0
12.0
50000 60000 70000 80000 90000 100000
Total System Inventory Count
D
e
l
i
v
e
r
y

M
e
a
n

T
a
r
d
i
n
e
s
s
DRP-200 DRP-600 DRP-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
50000 60000 70000 80000 90000 100000
Total System Inventory Count
D
e
l
i
v
e
r
y

M
e
a
n

T
a
r
d
i
n
e
s
s
OP-200 OP-600 OP-1000

-14-














Fig. 5 - Comparison with Lot Size Parameter k = 200













Fig. 6 - Comparison with Lot Size Parameter k = 600














Fig. 7 - Comparison with Lot Size Parameter k = 1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
50000 60000 70000 80000 90000 100000
Total System Inventory Count
D
e
l
i
v
e
r
y

M
e
a
n

T
a
r
d
i
n
e
s
s
DRP-600 OP-600
0.0
2.0
4.0
6.0
8.0
10.0
12.0
50000 60000 70000 80000 90000 100000
Total System Inventory Count
D
e
l
i
v
e
r
y

M
e
a
n

T
a
r
d
i
n
e
s
s
DRP-1000 OP-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
50000 60000 70000 80000 90000 100000
Total System Inventory Count
D
e
l
i
v
e
r
y

M
e
a
n

T
a
r
d
i
n
e
s
s
DRP-200 OP-200

-15-
Conclusions

A direct comparison of two common distribution planning and control systems, based on the
logic used to move material through supply chains, has been presented. Although there has been a lot
of conjecture regarding the relative merits of such systems, the current study is a step towards
understanding the true underlying behaviour and tradeoffs in each system. Results indicate that
centralized planning and control, as implemented under Distribution Requirements Planning (DRP),
is beneficial under realistic situations of time-varying demand and replenishment time uncertainty.
Further research is required to determine how different levels of variability and uncertainty
affect the relative performance of these systems. This research also does not address the selection of
proper planned lead times or reorder points on an a priori basis. It is also of interest to examine
analytical relationships which could guide in the selection of planning and control parameters based
on the uncertainty present and the desired level of delivery performance.
This research has focused on comparing the relative performance of systems based only on
the logic implemented. In practical terms, there are a number of additional issues related to
implementation. DRP systems require relatively complex and extensive software information
systems. As well, the various nodes in the supply chain or distribution network must all be able to
communicate with the central planning function. This may not be feasible if the network nodes are
controlled by independent enterprises. Less coordination of information is required when Order
Point systems are used. The use of only inventory and order information local to the upstream
replenishment loop makes implementation significantly easier. Such issues such concerning
implementation and operation may act to offset the advantages of systems when evaluated purely on
the basis of planning and control logic.

Acknowledgements

This research was supported by a grant from the Natural Sciences and Engineering Research
Council (NSERC) of Canada.


-16-
References

(1997). 1997 Benchmarking study shows dramatic reductions in costs using best practices for
supply-chain management. <http://www.PRTM.com/insight/insightarchier#SCM> (2000, June
15).
ARENA Users Guide, 1994, Sewickley, PA: Systems Modeling Corp.
Bookbinder, J. H., & Heath, D. B. (1988). Replenishment analysis in distribution requirements
planning. Decision Sciences, 19, 477-489.
Cooke, J.A. (1983). A dynamic approach to distribution. Traffic Management, 22(7), 55-58.
Handfield, R.B., & Nichols, E.L. (1999). Introduction to Supply Chain Management. New Jersey: Prentice
Hall.
Ho, C. (1990). Distribution requirements planning: A generalised system for delivery scheduling in a multi-
sourcing logistics system. International Journal of Physical Distribution and Logistics Management, 20(2),
3-8.
Ho, C. (1992). An examination of a distribution resource planning problem: DRP system nervousness. Journal
of Business Logistics, 13(2), 125-152.
How DRP helps warehouses smooth distribution. (1984). Modern Materials Handling, 39(6), 52-57.
Madia, R. D. (1990). DRP vs. order point inventory management: New solutions to old problems. Automation,
37(9), 44-46.
Martin, A. J. (1983). DRP (distribution resource planning)-Can you afford not to have it? Twenty-Sixth APICS
Annual Conference Proceedings. New Orleans, Louisiana, 52-55.
Maskell, B. H. (1988). Distribution resource planning: Part 2- A practical approach. Manufacturing Systems,
6(7), 64-69.
Ross, D. (1988). Integrating the information flow. Industrial Management, 12(9), 7-9.
Sipper, D., & Bulfin, R. L. (1997). Production Planning, Control, and Integration. New York: McGraw-Hill.
Smith, B. (1985). DRP improves productivity, profit, and service levels. Modern Materials Handling, 40(8),
63-65.