ECO A1

Value of Information in a Capacitated Supply Chain

Bhaswar Choudhury XLRI Jamshedpur, C.H. Area (East), Jamshedpur, India 831 001, Fax No. 00(91) (657)2227814, email: [email protected]

Yogesh K. Agarwal Indian Institute of Management, Lucknow Prabandh Nagar, Off. Sitapur Road, Lucknow, India 226 013, Fax No. 00(91)

(522)2734005, email: [email protected]

K. N. Singh Indian Institute of Management, Lucknow Prabandh Nagar, Off. Sitapur Road, Lucknow, India 226 013, Fax No. 00(91)

(522)2734025, email: [email protected]

D. K. Bandyopadhyay Indian Institute of Forest Management P.O. Box 357, Nehru Nagar, Bhopal, India 462 003, Fax No. 00(91) (755)2772878,

email: [email protected]

Abstract—The aim of this study is to understand the effect of end item demand variability, production flexibility, inventory characteristics, the number of retailers, and the impact of information sharing between channel members in a capacitated supply chain. Discrete event simulation is carried out to determine the impact of the above mentioned factors in a supply chain under different information sharing strategies. The simulation allows us to capturing the dynamic and stochastic complexity of the supply chain system. A two-level supply chain with one supplier having a capacity constraint and N identical retailers experiencing stationary and stochastic demand was built using a General Purpose Simulation Software. The optimum value of decision variables for various information scenarios is determined by multiple simulation runs and the response surface methodology technique.

Keywords Capacitated supply chain, information sharing, simulation analysis.

1. INTRODUCTION

Information technology has a significant impact on the supply

chain performance. It has radically influenced the supplier-

customer relationship by the possibilities introduced by the

abundance of data and the savings inherent in the analysis of

the data. It has been recorded that by exchanging information

related to inventory level, forecasting data, and sales trend,

companies are reducing their cycle times, fulfilling orders

more quickly, reducing excess inventory and improving custo-

mer service (Simchi-Levi and Zhao, 2003).

Sharing sales information has been suggested as an effective

strategy to reduce the impact of the bullwhip effect. In fact,

sharing demand information by the customer with his supplier

is the basis for initiatives like Quick Response (QR) and

Efficient Customer Response (ECR) (Lee et al., 2000).

Furthermore, information sharing is embedded in programs

like Vendor Managed Inventory (VMI) or Continuous

Replenishment Programs (CRP).

The objective of this simulation study is to test the belief that

the benefit is increased by sharing relevant information among

players in a supply chain. The paper studies the benefit to the

entire supply chain obtained by sharing information among

the members in the chain unlike previous work in the field

where the benefit was studied from the individual member’s

perspective. In addition to that, we investigate the underlying

drivers of the benefit and attempt to quantify the impact of

these drivers. For the purpose of this study, we consider three

degrees of information sharing. With Traditional Information

(TI) sharing, the information provided to the supplier is Received 21 November 2006, Revision 2 June 2008, Accepted

31 July 2008.

INFOR, Vol. 46, No. 2, May 2008, pp. 117–127 ISSN 0315-5986 j EISSN 1916-0615

117

limited to retailers’ replenishment order. In Retailer Managed

Inventory (RMI), the supplier has the knowledge of retailers’

daily inventory status and the demand experienced by the retai-

lers. But the retailers decide the order quantity based on the

stock on-hand and order-up-to level policy. Finally, in

Vendor Managed Inventory (VMI), the supplier has real time

information of the inventory status of the retailers. The supplier

assumes control of the stock management and is responsible for

deciding the quantity of shipments to the retailers. The optimal

order-up-to inventory level of retailers, and the start-up and

shut-down inventory levels of the supplier for TI policy are

determined by the simulation exercise to obtain minimum

system cost. In RMI and VMI, the decision variables are the

order-up-to inventory level and safety stock level of the retai-

lers, and the start-up and shut-down inventory level for the

suppliers.

The performance metric to measure the value of information

in the supply chain is the relative cost difference under TI and

RMI, and TI and VMI. The relative cost difference is the

maximum value of shared information. Optimum values of

decision variables for TI, RMI and VMI are determined by

multiple simulation runs. We use a simulation model because

of the dynamic and stochastic nature of the problem.

In the next section, the literature review is presented. The

methodology is described in Section 3. Section 4 describes

the model. Various operational parameters for the simulation

study are discussed in Section 5. In Section 6, we present the

results of the study. Finally, we conclude the discussion in

Section 7.

2. LITERATURE REVIEW

Forrester first proposed the concept of supply chain coordi-

nation using the theory of Industrial Dynamics (Forrester,

1961). He demonstrated how the delays, amplifications and

oscillations in the flow of demand information adversely

affect the supply chain.

It has been shown in a simple two-level system with one sup-

plier and one customer that the benefits are sensitive to demand

variability, the service level provided by the supplier and the

degree to which the order cycle of the customer and the pro-

duction cycle of the supplier are out of phase (Bourland et al.

1996).

The study of the information flow between a supplier and a

retailer in a two-echelon supply chain was conducted to under-

stand the relationships between capacity, inventory and the

information at the supplier level (Gavirneni et al. 1999). This

study also explored the impact of retailer’s ordering policy

and end item demand distribution on the supplier.

Lee et al. (2000) discussed the value of shared information

to improve the supply chain performance in a serial system

experiencing autoregressive demand. They showed that infor-

mation sharing lowered supply chain costs by about 23% in

their scenario with highest demand nonstationarity. Both

Gavirneni et al. (1999) and Lee et al. (2000) have studied the

value of information from the supplier perspective.

Cachon and Fisher (1997) examined forecasting and inven-

tory management under VMI. They found that both the retailer

and manufacturer inventories could be reduced while improv-

ing the service. The paper pays no attention to manufacturing

capacity constraints and inventory allocation among the

retailers.

Simchi-Levi and Zhao (2003) examined the frequency and

timing of demand information on the performance of the

supply chain in a simple two-stage supply chain with one sup-

plier and one retailer. Chen (1998) explored the relative benefits

of echelon-stock policies over those of installation stock pol-

icies in a multi-echelon environment. The paper discusses the

benefit of better communication between supply chain

members.

Waller et al. (1999) demonstrated that VMI reduces inven-

tory for all participants in the supply chain without adversely

impacting customer service level. Zhao et al. (2002) studied

the impact of a capacitated supply chain under different

demand conditions. Supply chain coordination is considered

with different levels of information sharing.

In our study, we consider one capacitated supplier and mul-

tiple retailers experiencing stationary and stochastic end item

demand. The supplier is the only source of inventory and

thus, there is no transshipment of inventory. The impact of

information sharing on the entire supply chain is considered

unlike Gavirneni et al. (1999) and Lee et al. (2000) where the

impact on the supplier is investigated. Also, they have con-

sidered a serial supply chain with one supplier and one retailer.

We have explored the supply chain with one supplier and N

identical retailers. We have also considered a capacitated

supplier unlike Cachon and Fisher (2000) where the supplier

has no capacity restriction. Waller et al. (1999) studied the

impact of TI and VMI on end item demand variability,

limited manufacturing capacity and partial adoption of VMI.

In our research, we examined TI, RMI and VMI, and its

impact on inventory characteristics and number of retailers,

in addition to end item demand variability and limited manu-

facturing capacity. We have not considered the partial adoption

of VMI. In our model, coordination is ensured by sharing

inventory information by the retailers whereas Zhao et al.

(2002) have ensured coordination by sharing forecast

information.

3. METHODOLOGY

A supply chain is a complex system where many factors are

operating simultaneously (Forrester, 1961). This makes it diffi-

cult to encompass all aspects of a real life system into a tra-

ditional mathematical model. So, to consider complex supply

chain network structures, more realistic demand structure

BHASWAR CHOUDHURY ET AL.118

INFOR, Vol. 46, No. 2, May 2008, pp. 117–127 DOI 10.3138/infor.46.2.117 ISSN 0315-5986 j EISSN 1916-0615 Copyright # 2008 INFOR Journal

necessitates the shift from simple analytical models to simu-

lation based research (Sahin and Robinson, 2002).

It was, therefore, decided to construct a discrete event simu-

lation model of a two echelon supply chain with one supplier

and N identical retailers experiencing stationary and stochastic

demand. For model building, we have used commercially avail-

able General Purpose Simulation System (GPSS). Discrete

event simulation allows us to evaluate a system as it evolves

through time. There is a queue of transactions that needs to

be processed during the simulation period. The logic that deter-

mines the processing of the transaction is embedded in the

simulation model. In this study, the logic considers local and

global inventory information depending on the level of the

information availability. Decision variables for the retailers

are the order quantity and the safety stock. Decision variables

for the supplier’s production process are the start-up and shut-

down level. These decision variables will be optimized for each

combination of parameters to make valid comparisons.

The length of the simulation run is selected in such a way

that the termination effect will be minimized. For the purpose

of the study, the model was run for the period of 73000 days

i.e., for each scenario, penalty cost of the retailers, holding

cost of the retailers, holding cost of the supplier and total

cost of the system were collected at the end of 73000 days.

The cost function is discussed in the next section.

4. MODEL DESCRIPTIONS

We consider 1 supplier and N identical retailers. The retailers

experience demands which are stationary and stochastic in

nature. The demand follows a normal distribution. The retailer

follows a periodic review inventory policy where the review

period is one week. The supplier is the only source of inventory

for the retailers. Thus, there is no transshipment or diversion of

stock. The supplier continuously reviews its inventory and

decides start-up and shut-down of production process based

on echelon/installation stock (depending on the information scenario). We assume negligible set-up time and set-up cost.

The sequence of events is as follows. (1) The supplier takes

the decision to set up its production process based on echelon/ installation stock. If production is set-up, the stock equivalent

to the production capacity is delivered to the supplier after 1

day production lead time. (2) The retailers experience daily

demand which is satisfied from the stock. Unfulfilled demand

results in lost sales (backorder is not considered). (3) The retai-

lers place orders to the supplier based on their order-up-to level

policy on every fifth day since the last replenishment ( for TI

and RMI). (4) Replenishment of an order is done on the

seventh day since the previous replenishment. The supplier, if

it has sufficient stock, entirely fulfills the retailers order. If

the supplier has less stock than the total demand of all the retai-

lers taken together, then the inventory is allocated among retai-

lers in proportion to the order placed by them (for TI and RMI).

For VMI, the supplier will gather the inventory status of each

retailer on a continuous basis and decide its production

start-up and shut-down processes. The replenishment of stock

for retailer is done on the seventh day from the previous replen-

ishment, in such a fashion that after replenishment, all the retai-

lers will carry the same quantity of stock. (5) Inventory holding

cost and stock out cost is charged at the retailer and carrying

cost is accounted for at the supplier.

Demands not filled immediately are lost sales and a penalty

cost is charged for the same. There are holding and penalty

costs for each day. Let hs . 0 be the per unit per day inventory

holding cost of stock at the supplier and hr . 0 be the per unit

per day inventory holding cost of stock at the retailer, respect-

ively. We assume hr ¼ 2 hs. Let pr be the per unit penalty cost

due to stock out. Let E[C] be the expected per period supply

chain holding and stock-out ( penalty) costs for a given inven-

tory policy followed by the retailers and the supplier. We will

use notation x þ

to denote max (0, x). The cost of the system

is defined by the following equation:

E½C� ¼ hsE½Ies �þSN½hr EðI e ri � yÞ

þ þ pr Eðy � IeriÞ þ� ð1Þ

where E[I e s] is the expected inventory with the supplier at the

end of the day, N is the number of retailers considered in the

system, E(I e ri 2 y)

þ is the expected inventory with the retailer

i at the end of the day after experiencing demand y,

E( y 2 I e ri) þ

is the expected stock out experienced by the retai-

ler i at the end of the day after experiencing demand y. The time

discount factor is assumed to be 1.

For the purpose of the study, we evaluate the benefit due to

information sharing by the following equation.

B ¼ ðTI � MIÞ ðTlÞ

ð2Þ

where B ¼ % Benefit due to information sharing, TI ¼ Total

supply chain cost for Tradition Information Policy,

MI ¼ Total supply chain cost for Modified Information

Policy, and the modified information policies are RMI or VMI.

4.1 TI Policy Model

In this model, the supplier has no information about the retai-

lers’ inventory status or the demand experienced by the retai-

lers. The supplier experiences periodic replenishment orders

from the retailers. This is illustrated as orders originating

from a black box (retailer) in Fig. 1. Retailers experience iden-

tical and independently distributed demand (i.i.d).

The retailer practices order-up-to level policy. On every fifth

day since the previous replenishment, the retailer places an

order to the supplier based on her inventory policy. The sup-

plier replenishes the retailer on the seventh day since the pre-

vious replenishment. The supplier decides its production

process based on the local inventory position. This is akin to

a reorder point policy. If the stock is available with the supplier,

VALUE OF INFORMATION IN A CAPACITATED SUPPLY CHAIN 119

INFOR, Vol. 46, No. 2, May 2008, pp. 117–127 DOI 10.3138/infor.46.2.117 ISSN 0315-5986 j EISSN 1916-0615 Copyright # 2008 INFOR Journal

then the supplier honors the retailers’ demand. If the stock is

not sufficient then the supplier allocates the inventory among

the retailers in proportion to their order.

Thus, in TI, the production start-up and shut-down decision

can be defined as:

If supplier’s stock � supplier’s start-up level of inventory; then start production;

If supplier’s stock � supplier’s shut-down level of inventory, then stop production:

The shipment decision is defined as:

If supplier’s stock � Si retail orderi; then ship to each retailer their entire order;

OR

If supplier’s stock , Si retail orderi, then ship to each retai-

ler in proportion to the order placed by the retailers, and suppli-

er’s stock ¼ 0.

The optimum value is obtained by carrying out multiple

simulation runs to get a minimum value for (1) by optimizing

the order up-to level for the retailers, and the start-up and shut-

down inventory levels for the supplier.

4.2 RMI Policy Model

In this scenario, illustrated in Fig. 2, the retailer communicates

daily inventory status to the supplier. The retailer practices

order-up-to level policy. On every fifth day since the previous

replenishment, the retailer places an order to the supplier

based on its stock and inventory policy. The supplier

replenishes the retailer on the seventh day since the last replen-

ishment. Since the supplier has the visibility of the retailers’

stock; it decides its production start-up and shut-down based

on the echelon stock (the echelon stock of echelon i is the

number of units in the system that are at, or have passed

through, echelon i, but have not been specifically committed

to outside customers). If enough stock is available with the sup-

plier, it fulfills the retailers demand. If sufficient stock is not

available, then the inventory is allocated among the retailers

in proportion to the order placed by them.

Thus, in RMI, the production start-up and shut-down

decisions can be defined as:

If echelon stock of supplier � supplier’s start-up inventory; then start production,

If echelon stock of supplier . supplier’s shut-down

inventory, then stop production.

A parameter safety inventory is introduced for echelon

inventory calculation, which is optimized in the simulation

process. The logic for the calculation of echelon inventory is

as follows: If, in a supply chain, the following conditions are

met: (1) the stock at any retail outlet (SRi) is much higher

than the safety inventory (SIi); (2) the stock at any other

retail outlet (SRj) is much less than the safety inventory (SIj);

and (3) the aggregate stock at all retail outlets (SiSRi) and at

supplier’s level is greater than or equal to the echelon stock

shut down inventory level; then, the excess inventory at the

retail outlet (SRi) may not be able to protect the supply chain

from experiencing a stock-out because of low inventory level

at the retail outlet (SRj). Thus,

If stock RiðSRiÞ , safety inventoryðSIiÞ; then stock Ri is considered for the calculation of echelon inventory;

If stock RiðSRiÞ� safety inventoryðSIiÞ; then safety inventory is considered for the calculation of echelon stock.

The shipment decision is defined as:

If supplier’s stock � Si retail orderi; then ship to each retailer their entire order;

OR

If supplier’s stock , Si retail orderi; then ship to each retailer in proportion to the order placed by all retailers;

and supplier’s stock ¼ 0:

Figure 2. Retailer Managed Inventory PolicyFigure 1. Traditional Information Policy

BHASWAR CHOUDHURY ET AL.120

INFOR, Vol. 46, No. 2, May 2008, pp. 117–127 DOI 10.3138/infor.46.2.117 ISSN 0315-5986 j EISSN 1916-0615 Copyright # 2008 INFOR Journal

The optimum value is obtained by carrying out multiple simu-

lation runs to get a minimum value for (1) by optimizing the

order up-to level, the safety stock for the retailers, and the

start-up and shut-down levels for the supplier.

4.3 VMI Policy Model

In this scenario, the supplier has the visibility of retailers stock

on a continuous basis. So, it can coordinate its production

start-up and shut-down based on the echelon stock. The pro-

duction process is identical to that of RMI. The difference

between RMI and VMI policies is in the inventory allocation

to the retailers. In VMI, if the supplier has a stock more than

the aggregate difference in order-up-to level and stock

on-hand for all retailers, then the supplier will replenish them

up to the order-up-to level. If the supplier has a stock less

than the aggregate difference in order-up-to level and stock

on-hand for all retailers, then the supplier will replenish them

in such a fashion that each retailer will have the same stock

after replenishment. The supplier replenishes the retailers on

the seventh day since the last replenishment.

Thus the shipment decision of the supplier in VMI is defined

as:

If supplier’s stock � Siðorder-up-to level � stock RiÞ; then fill up to the order-up-to level;

OR

If supplier’s stock , Siðorder-up-to level � stock RiÞ; then fill up to the level such that; after replenishment; stock

Ri ¼ stock Rj for i = j 8 i; j and supplier’s stock ¼ 0:

The optimum value is obtained by carrying out multiple

simulation runs to get minimum value for (1) by optimizing

order-up-to level and safety stock of the retailers and the

start-up and shut-down levels for the supplier.

5. SIMULATION PARAMETERS

Information sharing and coordination between channel partners

is a commonly suggested solution to mitigate inefficiencies in a

supply chain. The review of the literature on the subject

suggests a number of factors that have a bearing on the perform-

ance of the supply chain. A thorough understanding of

the influence of these factors is critical to decision makers.

The simulation model facilitates the study of the impact of

the different parameters on the performance of the

supply chain.

5.1 Demand Variability

Demand variability is the most widely recognized factor that

has the potential to influence the supply chain performance

(Gavirneni et al., 1999; Waller et al., 1999, Zhao et al.,

2002). Researchers have shown in their models that as the

demand variability increases, the percentage benefit due to

availability of information drops (Chen, 1998; Gavirneni

et al., 1999). Since the finding is counter-intuitive, further

investigation is needed to explore the impact of demand varia-

bility on the performance of a supply chain. For each of the

information sharing and coordination scenarios (TI, RMI and

VMI), we specify five levels of demand variability faced by

the retailers. The variability was measured by the coefficient

of variation (CV). Thus, in our system, we consider low

demand variability (CV ¼ 0.05) to high demand variability

(CV ¼ 0.25).

So, for four retailers in a supply chain network, we consider

the following demand patterns:

�Nð400; 100Þ;�Nð400; 80Þ;�Nð400; 60Þ; �Nð400; 40Þ and �Nð400; 20Þ:

5.2 Inventory Holding Cost and Stock-Out Cost

We study the impact of inventory holding cost and stock-out

cost on the optimal policy under TI, RMI and VMI scenarios.

We assume the selling price of the product is 100 Rupees

(Rs) per unit. Assuming a 20% profit margin, let the stock

out cost be 20 Rs per unit. Assuming inventory carrying cost

of 20% per annum, then the holding cost per unit per day is

approximately 0.055 Rs f(100 � 20/100)/365g. Hence, the following holding costs are considered: 0.03 Rs per unit per

day, 0.06 Rs per unit per day and 0.10 Rs per unit per day to

study its influence on the total system cost under the three infor-

mation sharing and flow coordination scenarios.

5.3 Supplier’s Capacity Tightness

This refers to how tight is the supplier’s production capacity

relative to the demand. It is defined as the ratio of total

demand to the total capacity available to satisfy the demand.

Capacity tightness is an indicator of supplier’s flexibility. A

number of authors have investigated the effect of flexibility

on supply chain responsiveness (Cachon and Fisher, 1997;

Gavirneni et al. 1999, Simchi Levi and Zhao, 2003) Eight

levels of capacity tightness are considered, i.e., Demand/ Capacity (D/C) ¼ 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, where 0.9 corresponds to very high capacity tightness and 0.2 corre-

sponds to very low capacity tightness.

5.4 Number of Retailers

The effect of the number of retailers involved in a supply chain

network is an important consideration. As the size of the retail

network grows, the pooling of variability at the supplier level

increases. But, the system inventory is more fragmented and

distributed across a large number of retailers. A thorough

understanding of this factor is necessary.

The impact of information with change in the number of

retailers (N) is studied at 9 levels. Those are 1, 2, 4, 6, . . .,

VALUE OF INFORMATION IN A CAPACITATED SUPPLY CHAIN 121

INFOR, Vol. 46, No. 2, May 2008, pp. 117–127 DOI 10.3138/infor.46.2.117 ISSN 0315-5986 j EISSN 1916-0615 Copyright # 2008 INFOR Journal

and 16 retail network systems. The suppliers capacity is con-

stant for all the network system with D/C ¼ 0.6. Waller et al. (1999) has studied the system with seven retail networks,

but considered priority allocation of inventory to a selected

subset of retailers [14]. Zhao et al. (2002) has explored a four

retail network system with direct shipment practice [15].

In our two echelon supply chain model, each retailer experi-

ences identical and independently distributed (i.i.d) daily

demand. Total system demand is 1600 units per day. Mean

demand at each retail level would be 1600/n, where n is the number of retailers in the system. This is to avoid confounding

the effect of increasing the system demand and the number of

retailers proportionately when one needs to explore the impact

of the number of retailers on the system. Standard deviation is a

parameter. Thus, the demands experienced by the system

under different configurations are given in Table 1.

6. RESULTS

In this section, our goal is to study the trade-offs between the

end item demand variability, the inventories, the capacities,

the number of retailers and the information. We vary the end

item demand variability, the inventory holding cost, the

capacity, and the number of retailers. For each scenario, the

system-wide cost-minimizing inventory policy for the retailer

and the supplier is determined by simulation.

6.1 The Effect of Demand Variability

The plot of benefit (in percent) versus end item demand var-

iance between the TI and the VMI, and the TI and the RMI,

is shown in Fig. 3 (a) and Fig. 3 (b), respectively. It is observed

that the benefit due to availability of information to the system

is higher when the coefficient of variation is high than when it

is low. This is because the average inventory includes two com-

ponents: one proportional to average periodic demand and the

other proportional to the standard deviation of the periodic

demand (safety stock). The reduction in inventory is achieved

through the reduction in safety stock. With the availability

of real time information, the uncertainty associated with

the demand variability reduces. Hence, the higher the coeffi-

cient of variation, the greater is the impact of information in

reducing the uncertainty, the higher is the reduction in the

safety stock.

It is also observed from Fig. 3 (a) and Fig. 3 (b) that the

benefit to the supply chain is always higher in the VMI scenario

in comparison to the RMI scenario. This is because the inven-

tory allocation among retailers by the supplier is more effective

in the VMI scenario than in the RMI scenario.

TABLE 1

Demand Distributions with Respect to the Number of Retailers

Number of retailers Demand distribution

1 N(1600,400)

2 N(800,200)

4 N(400,100)

6 N(266.66,66.67)

8 N(200,50)

10 N(160,40)

12 N(133.33,33.33)

14 N(114.28,28.57)

16 N(100,25)

Figure 3. (a) Plot of % Benefit versus Coefficient of Variation between TI and VMI

BHASWAR CHOUDHURY ET AL.122

INFOR, Vol. 46, No. 2, May 2008, pp. 117–127 DOI 10.3138/infor.46.2.117 ISSN 0315-5986 j EISSN 1916-0615 Copyright # 2008 INFOR Journal

6.2 The Effect of Inventory Holding Cost

The Figures below show the plot of the total supply chain

cost versus the holding cost. Fig. 4(a) and Fig. 4(b) is for

different coefficient of variation for the TI and the VMI,

and the TI and the RMI respectively. It is observed that as

the holding cost increases, the total system cost increases.

But, with the increase in holding cost per unit per day, the

benefit due to the availability of information in terms of

cost saving increases. The availability of information helps

in rationalizing the inventory in the supply chain network.

The saving in terms of absolute cost is more as the holding

cost increases.

Another observation is that as the holding cost increases,

the relative benefit decreases. This is shown in Fig. 5 (a)

and Fig. 5 (b). This can be explained as follows: if the

holding cost is low, it implies that the relative penalty cost

is high. As mentioned before, the penalty cost in our model

is kept constant and the holding cost is changed. This is

because proportionally changing both penalty and holding

costs does not change the results, and it is sufficient to vary

Figure 4. (a) Plot of Total System Cost versus Inventory Holding Cost for Different Coefficient of Variations for TI and VMI

Figure 3. (b) Plot of % Benefit versus Coefficient of Variation between TI and RMI

VALUE OF INFORMATION IN A CAPACITATED SUPPLY CHAIN 123

INFOR, Vol. 46, No. 2, May 2008, pp. 117–127 DOI 10.3138/infor.46.2.117 ISSN 0315-5986 j EISSN 1916-0615 Copyright # 2008 INFOR Journal

only one of them (Gavirneni et al. 1999). Thus, the system

derives the benefit from the availability of information by

avoiding stock outs.

Thus, one may conclude from the above observation that the

role of inventory holding cost for the benefit of the supply chain

depends on the perspective of the decision maker. When rela-

tive benefit is the matter of concern, then the benefit to the

supply chain is most at low inventory holding cost (Rs 0.03

per unit per day). But, if the absolute benefit is of prime import-

ance to the manager then the savings due to the information

sharing increases with an increase in inventory holding cost

(0.10 Rs per unit per day).

6.3 The Effect of Capacity Tightness

To study the value of information sharing as a function of pro-

duction capacity, we illustrate in Fig. 6 (a) and Fig. 6 (b), the

percentage benefit derived in terms of cost saving from VMI

and RMI over TI, respectively.

Eight different levels of capacity tightness and three differ-

ent levels of end item demand variance are considered for the

study.

The analysis demonstrates that as the production capacity

increases, the benefit, in terms of percentage cost saving,

increases. This can be explained as follows. As the production

capacity increases, the optimal policy would be to postpone

Figure 4. (b) Plot of Total System Cost versus Inventory Holding Cost for Different Coefficient of Variations for TI and RMI

Figure 5. (a) Plot of % Benefit versus Inventory Holding Cost between TI and VMI

BHASWAR CHOUDHURY ET AL.124

INFOR, Vol. 46, No. 2, May 2008, pp. 117–127 DOI 10.3138/infor.46.2.117 ISSN 0315-5986 j EISSN 1916-0615 Copyright # 2008 INFOR Journal

production as much as possible and take advantage of all the

information available before committing production. For

example, if infinite capacity is available, it is optimal to wait

until the last bit of information is available. On the other

hand, if there is a capacity constraint, then the production com-

mitment is determined by the capacity.

Another observation is that beyond some value of the

capacity tightness (Demand/Capacity ¼ 0.3 in present case), the percentage benefit derived due to the availability of infor-

mation becomes steady. This is because at high slackness in

capacity, the production initiation is delayed until the last

possible information is available. The marginal utility of

capacity to delay production is negligible, beyond a particular

ratio of demand to capacity.

By studying the plots of TI and RMI, we notice that at

Demand to Capacity (D/C) ratio of 0.5 and less, the supply chain cost is insensitive to change in end item demand var-

iance. But, the same phenomenon is not visible in the plots

of TI and VMI. This implies, that at D/C ¼ 0.5 and less, the production coordination due to real time visibility of infor-

mation is not important. But, the allocation of supplier’s

stock among retailers becomes relevant.

Figure 5. (b) Plot of % Benefit versus Inventory Holding Cost between TI and RMI

Figure 6. (a) Plot of % Benefit versus the Ratio of Demand to Capacity between TI and VMI

VALUE OF INFORMATION IN A CAPACITATED SUPPLY CHAIN 125

INFOR, Vol. 46, No. 2, May 2008, pp. 117–127 DOI 10.3138/infor.46.2.117 ISSN 0315-5986 j EISSN 1916-0615 Copyright # 2008 INFOR Journal

6.4 The Effect of the Number of Retailers

Fig. 7 illustrates the impact of information sharing as a function

of the number of retailers in the supply chain system. For the

purpose of this study, we have kept the end item demand

variance (coefficient of variation) constant for all the retailers.

It is observed that as the number of retailers in the supply

chain increases, the value of information diminishes. This

can be explained as follows: as the demand is aggregated

from different locations, it becomes more likely that the

higher than average demand from one customer will be

offset by the lower than average demand by the other customer.

Thus, in the base case (Traditional Information), the total

system cost is reduced. So, the benefit of additional information

diminishes when the number of retailers is increased.

Figure 6. (b) Plot of % Benefit versus the Ratio of Demand to Capacity between TI and RMI

Figure 7. Plot of % Benefit versus the Number of Retailers

BHASWAR CHOUDHURY ET AL.126

INFOR, Vol. 46, No. 2, May 2008, pp. 117–127 DOI 10.3138/infor.46.2.117 ISSN 0315-5986 j EISSN 1916-0615 Copyright # 2008 INFOR Journal

7. CONCLUSIONS

In this simulation study, we considered a two-stage supply

chain with multiple retailers facing i.i.d demand and a single

manufacturer with finite production capacity. We considered

three different information sharing policies: TI, RMI and

VMI. By analyzing the model, we studied the benefit derived

by the entire supply chain due to sharing of demand and inven-

tory information with the supplier.

It is observed that as the end item demand variance

increases, the benefit due to the availability of real time inven-

tory information increases. This is contrary to the previous

studies carried out by other researchers (Chen et al. 1998;

Gavirneni et al., 1999).

It is shown that when the relative benefit is of paramount

importance, then a low inventory holding cost is beneficial to

the supply chain. But, the saving increases with an increase

in inventory holding cost if the absolute benefit is of prime

importance to the supply chain manager.

We have observed that the potential benefit of information

sharing between channel members increases as the supplier’s

capacity increases. But, beyond a particular ratio of Demand

to Capacity (D/C), the marginal utility of each additional capacity is not substantial. It is also observed that the additional

benefits due to the coordination of production, are insensitive to

end item demand variance considered in the study for D/ C ¼ 0.5 and less. At this point, the allocation of inventory by

the supplier’s among retailers reaps benefits. In other words,

there is a critical D/C ratio beyond which the RMI policy is indifferent to end item demand variance, but that need not be

the case for the VMI policy.

This study illustrates that as the retail network grows, the

benefit to the supply chain with the availability of real time

information reduces. Finally, the VMI scenario is more ben-

eficial than the RMI scenario in all cases. This implies inven-

tory allocation among channel partner under information

sharing condition has greater impact in cost reduction than

only coordination between players by information sharing.

The study supplements the research corroborating positive

impact of information technology on the business performance.

The model considered is representative of many supply chain

situations, but the conclusion is limited to the context we

have explored.

In the model, we assume there is no conflict of interest

among channel members in sharing relevant information.

Second, we considered each retailer is identical and experien-

cing i.i.d demand. In our model, we have considered negligible

set-up time and set-up cost. We intend to relax these constraints

in our future work with suitable empirical data. Finally, in our

model the retailer practices order-up-to policy with constant

order-up-to level. This assumption implies that the retailer is

not able to anticipate future shortages at the supplier. If the

retailer anticipates shortage, then its rational action would be

to inflate the order. Thus, the behavior of rationing and

gaming is not captured in our model.

ACKNOWLEDGEMENTS We would like to express our sincere thanks to the two blind

referees for their invaluable suggestions.

REFERENCES [1] Bourland, K.E., Powell, S.G., and Pyke, D.F. (2000), “Exploiting

Timely Demand Information to Reduce Inventories”, European

Journal of Operational Research, 92 (1996): 239–253.

[2] Cachon, G. and Fisher, M. (1997), “Campbell Soup’s Continuous

Replenishment Program: Evaluation and Enhanced Inventory

Decision Rules”, Production and Operation Management,

6 (3): 266–276.

[3] Cachon, G. and Fisher, M. (2000), “Supply Chain Inventory

Management and the Value of Shared Information”,

Management Science, 46 (8): 1032–1048.

[4] Chen, F. (1998), “Echelon Reorder Points, Installation Reorder

Points, and the Value of Centralized Demand Information”

Management Science, 44 (12): S221–S234.

[5] Fisher, M.L. (1997), “What is the Right Supply Chain for Your

Product?” Harvard Business Review, 75 (2): 105–116.

[6] Forrester, J.W. (1961), Industrial Dynamics, The MIT Press,

Cambridge, Massachusetts.

[7] Gavirneni, S., Kapuscinski, R., and Tayur, S. (1999), “Value of

Information in Capacitated Supply Chains”, Management

Science, 45 (1): 16–24.

[8] Ghiani, G., Laporte, G., and Musmanno (2004), Introduction to

Logistics Systems Planning and Control, John Wiley & Sons,

Ltd.

[9] Hammond, J.E. and Barilla, SpA (1994), Harvard Business

School Case Study, 9-694-046, Harvard University,

Cambridge, MA.

[10] Lee, H.L., So, K.C., and Tang, C.S. (2000), “The Value of

Information Sharing in a Two-Level Supply Chain”,

Management Science, 46 (5): 626–643.

[11] McKenney, J.L. and Clark, T.H. (1995), “Campbell Soup Co.: A

leader in Continuous Replenishment Innovation”, Case 9–125–

124, Harvard Business School Publishing, Boston, MA.

[12] Sahin, F. and Robinson, E.P. (2002), “Flow Coordination and

Information Sharing in Supply Chains: Review, Implications

and Direction for Future Research”, Decision Sciences, 33 (4):

505–536.

[13] Simchi-Levi, D. and Zhao, Y. (2003), “The Value of Information

Sharing in a Two Stage Supply Chain with Production Capacity

Constraints”, Naval Research Logistics, 50: 888–916.

[14] Waller, M., Johnson, E., and Davis, T. (1999), “Vendor Managed

Inventory in the Retail Supply Chain”, Journal of Business

Logistics, 20 (1): 183–202.

[15] Zhao, X., Xie, J., and Zhang, W.J. (2002), “The Impact of

Information Sharing and Ordering Coordination on Supply

Chain Performance”, Supply Chain Management: An

International Journal, 7 (1): 24–40.

VALUE OF INFORMATION IN A CAPACITATED SUPPLY CHAIN 127

INFOR, Vol. 46, No. 2, May 2008, pp. 117–127 DOI 10.3138/infor.46.2.117 ISSN 0315-5986 j EISSN 1916-0615 Copyright # 2008 INFOR Journal