Impact of Information and Communication Technologies on Logistics ...

■ Transportation Research Record 1790 Paper No. 02- 2913

10

Impact of Information and Communication Technologies on Logistics and Freight Transportation Example of Vendor-Managed Inventories Mouhamad Rabah and Hani S. Mahmassani A framework to assess the impacts of information and communication technologies (ICTs) on supply-chain and logistics operations, including the movement of freight, is presented. The framework reflects transformation of the traditional supply chain into a virtually forged one and its link to the transportation sector. To illustrate and quantify cost savings that might result from information sharing among supply-chain participants, the example of a vendor-managed-inventory (VMI) strategy in a two-echelon distribution channel is considered. Models of operation with and without VMI are developed; the results of applying the two strategies to the same demand-stream characteristics allow assessment of the impact of ICT-enabled VMI logistics strategy. Reductions in total cost are achieved across all cases considered.

According to the Council of Logistics Management, “logistics is the process of planning, implementing, and controlling the efficient, effective flow and storage of goods, services, and related information from point of origin to point of consumption for the purpose of conforming to customer requirements” (1). In recent years, coupled with the developments in information and communication technologies (ICTs), logistics operations have evolved into supply-chain management. According to Metz, “Integrated Supply Chain Management (ISCM) is a process-oriented, integrated approach to procuring, producing, and delivering products and services to customers” (2). The scope of ISCM includes subsuppliers, suppliers, internal operations, trade customers, retail customers, and end users, and it covers the management of materials, information, and fund flows. The major factor that affected business logistics during the 1980s was the move toward internal integration, accompanied by the emergence and growth of third-party logistics firms. The main contributions of ICT were the compression of the supply chain and the increase in information trading to reduce inventory. These led to the emergence of new production and logistics strategies, such as the distribution-resource-planning (DRP) and the just-in-time (JIT) strategies. In the 1990s, the main forces that shaped business logistics included • Globalization, • Demographic forces, • ICT,

M. Rabah, Department of Civil Engineering, and H. S. Mahmassani, Department of Civil Engineering and Department of Management Science and Information Systems, University of Texas at Austin, Austin, TX 78712.

• Competition, • Government regulations, and • Environmental considerations. First, globalization is noticeable in the greater role that international logistics and third parties with specialized skills play. Second, companies must cope with consumers demanding much higher levels of service, and they are turning to outsourcing as a way to cope with possible labor turnover. Third, the main feature of ICT in the 1990s was the Internet, which is allowing direct buying, and the emergence of powerful modeling tools to exploit the wealth of consumer and business data in the supply chain. Fourth, due to increased competition, supply-chain optimization is necessary. Finally, government regulations are allowing new cross-border and trade agreements that are dramatically changing the operational landscapes of operating firms (3). ICT offers tools that improve the operational performance of individual functions such as monitoring and tracking devices. They also increase customer satisfaction through tools that enable easy ordering and status checking. In addition, they facilitate integration across the supply chain, and enable strategic alliances between shippers, carriers, and carrier users. Finally, they allow new business strategies and new channels to customers. Some of the new logistics strategies enabled by ICT include (a) vendor-managed-inventory (VMI) strategy, through which the supplier knows when and how much of its products to store at a major distributor or retailer; (b) merge-intransit (MIT) policy, with which the carrier or the third-party provider assembles the demanded product for final delivery from its various incoming components; (c) time-definite delivery (TDD), which ensures synchronous deliveries to multiple destinations by managing transport providers, service levels, and shipment release time; and (d ) the freeze-point delay, in which the final product assembly is left as close as possible to the point of end-consumption. Many researchers have investigated the impact of ICT on supplychain processes and highlighted the importance of collaboration in order to improve operations. D’Amours et al. analyzed the impact of information sharing in a make-to-order network manufacturing operation, based on three bidding protocols (4). Firms were supposed to schedule the required operations through a virtual manufacturing and logistics network (i.e., manufacturing, storage, and transportation), with the objective of minimizing cost so as to deliver the products in an allowed time window. Wang presented a conceptual framework for virtual markets accounting for “meta-level facilities,” “constraint enforcement,” and “inference facilities” (5). Lancioni et al. analyzed the use of the Internet in logistics operations across the supply chain,

Rabah and Mahmassani

Paper No. 02- 2913

including transportation, inventory management, warehousing, and vendor relations (6). The study was based on a random sample of 181 participants at the 1998 Council of Logistics Management conference in Los Angeles. Results showed that Internet use in the supply chain was mostly adopted (56.2%) in managing the firms’ transportation systems. Order processing was second, followed by “relations with vendors” (6). Inventory management ranked fifth. As mentioned, technological developments have altered fulfillment strategies, requiring faster, more reliable, and less expensive delivery policies. These policies are intended to help companies achieve better customer-level service, at lower costs. Moreover, developments in electronic commerce have been taking place at a rapid pace, in a dynamically changing environment. As researchers continue to grapple with its likely impacts, no definitive comprehensive framework has been successfully elaborated to assess the impacts of electronic commerce on supply-chain and logistics operations, including the movement of freight. Developing this framework is a primary motive for this research. Finally, Delaney reports that American companies, during 1997, spent $862 billion (10% of U.S. gross national product) on supply activities, including “the cost of movement, storage, and control of products across the supply chain” (7 ). Quantifying cost reductions in logistics operations through information technologies is a major motivation for this research. The rest of the paper is organized as follows: a discussion of the advancement in information technologies, including presentation of the interrelation between the Internet sector and transport activities; an analysis of the impacts of ICT on logistics relational policies and freight activity; a focus on cost reductions achievable through information sharing when adopting a VMI strategy in a two-echelon distribution channel; and finally, a summary of the findings of this work and an outline of future research directions.

RELEVANT ADVANCES IN ICT Advances in performance, decreases in prices of information systems and computers (Figure 1), expanded communication capabilities, wider availability of powerful software, and easier access to databases have accompanied the widespread use of computers and information technology (IT).

120 100

Actual

Trend decline of 26.2 percent between 94Q4 and 99Q4

50 25 Trend decline of 12.1 percent between 87Q1 and 94Q3

10 5 1987

1989

1991

1993

1995

1997

1999

FIGURE 1 Price declines in computers allow greater use of information technologies. (Note: log scale; index 1987 Q1 = 100. Source: U.S. Bureau of Economic Analysis.)

11

In the logistics area, applications focus on information to forecast usage; control replenishment: efficiently operate warehouses and shipping terminals; organize and schedule vehicle operations; and manage inventories, procurement, and material flows. These advances have motivated the development of powerful software and decisionsupport systems to help companies in the following areas (8): • Logistics network design; • Inventory deployment and management; • Distribution management; • Fleet management; • Production planning; • Demand planning and forecasting; • Procurement and purchasing (such as managing supplier databases; sending bid requests; and tracking information, order, and shipment status); • Warehouse control system and transportation management that improves customer-service performance; and • Information sharing, including data on order entry, requisitions, and bills of material. To maintain a competitive market position, companies have sought to optimize operations while meeting customers’ higher service-level expectations. These higher levels of service could be achieved through the adoption of new information technologies (software), information sharing, collaboration, and outsourcing some internal functions that have altered the traditional supply chain. Traditional flow channels were composed of linear exchanges of information and goods among the various parties that form the links of the chain. Collaboration across the supply chain was limited, and end-customers would obtain their products from retailers. ICT and specifically the Internet are allowing end-customers, whether individuals or firms, to order electronically from any collaborating party. Advances in technology and the Internet have created new channels for information flow, goods flow, and new partnerships across the supply chain that are affecting logistics and transport activity (Figure 2). As shown in Figure 2, the Internet plays the role of a single point of contact facilitating information exchange and opening new channels for goods flow. It is a form of a client-server computing platform, through which information is retrieved from ports distributed all over the world. Technologies that are allowing information exchange along these new channels include (a) e-mail, (b) online order forms and documents, (c) electronic payment, (d ) electronic-data interchange (EDI), and (e) enterprise resource planning (ERP), transactions that integrate manufacturing, financial, and other systems (9). Choice among these technologies depends on cost of implementation and the goals set by contributing firms. Internet tools (such as e-mail, online order forms, and electronic payments) offer the cheapest way of information exchange and are helping small firms become more competitive. The Internet contributors are divided into five categories: • Electronic retailers. These channels focus on the business-toconsumer (B2C) sector. Sales at these sites mainly substitute for the conventional sector. They provide all types of consumer commodities. Some of them are virtual entities, whereas others have a physical presence also. In the last two years, many bricks-and-mortar retailers have expanded their operations by moving online. These parties are regaining market share from e-tailers, many of whom where forced to shut down due to infeasible business models, among other reasons. An example of an e-tailer is Amazon.

12

Paper No. 02- 2913

Transportation Research Record 1790

VMI Strategy

Raw Material Supplier

Manufacturer

Wholesaler

Transport Provider

Internet

- Own Transporter - US Postal Service - Common Carriers - Third Party Logistics (3PL)

Order Tracking: Bar Coding

Home End Deliveries

Reverse Logistics

Retailer

Companies Demand

- Electronic Retailers - Services & Catalogs - Exchanges - Procurement Hubs - Auctions

Electronic Goods

Customers Demand

Reverse Logistics

Goods Flow Information Flow FIGURE 2

Transformation of the traditional supply chain into a virtually connected one.

• Services and catalogs. These sites act as intermediaries, providing information and digitizing catalogs for firms. They facilitate the shopping experience for customers by providing information sharing, advertisements, and cost comparison among products. The main revenue stream for these sites is online advertising, a sector that encountered a slowdown during 2000–2001, but it is still experiencing a high growth rate. An example in this category is Yahoo. • Exchanges. The objective of these sites is to help different parties meet and settle on prices for negotiated commodities by providing explicit information about the products. Currently, all types of commodities are being exchanged, and sites have various rules governing such exchanges. These sites help companies reduce or clear unwanted inventories by selling them at reduced prices. An example here is eBay. • Procurement and collaboration hubs. These sites are mainly involved in business-to-business (B2B) relations. Procurement hubs are clusters of organizations that have set their procurement site over the Internet; they attract suppliers to join in. Collaboration hubs not only offer a transaction medium but also help participating partners complete their projects from the design phase to the distribution stage. B2B marketplaces are growing fast and are likely to affect logistics operations by increasing the operational efficiency of collaborating parties and speeding up reactions to market changes. • Auctions. These sites form an evolving class of sales channels for many industries, including those selling consumer items. These could be buyer or seller oriented. Sites include Adauction and EWanted, a reverse auction site where sellers bid down their price to

make the sale. Many marketplaces are auctioning truck capacities, freight movement, inventory surplus, and inventory storage. These auctioning sites are likely to affect logistics operations by improving capacity usage and allowing better information sharing. Examples of leading sites include nte.net, logistics.com, and transplace.com. These five categories affect strategic relations among various parties and freight movement, subjects discussed in the following section.

IMPACT ON LOGISTICS RELATIONS AND FREIGHT MOVEMENT Channels representing goods flow are affected by the electronic economy. Traditionally, goods were shipped, in sequence, from the supplier to the manufacturer to the wholesaler to the retailer and finally to the customer. As shown in Figure 2, demands could be directly satisfied by any of the participating parties. Manufacturers, wholesalers, and retailers could be directly selling to individual or corporate customers through the Internet. Moreover, we are seeing several relationships developing across the supply chain. Types of coordination that are improving firms’ operational planning include • Buyer-vendor relations. Collaborating parties invest in material handling through data technologies such as EDI. This type of relation tries to achieve higher levels of customer service by decreasing


Paper No. 02- 2913

stock-outs while cutting inventory costs; moreover, collaborating parties try to improve demand-forecasting strategies. An example of such relations is VMI. • Production-distribution relations. This type of coordination helps collaborating parties determine vehicle routing, machine scheduling, and inventory buffers. These relations increase the efficiency of distributing to distribution centers (DCs), retailers, plants, and sometimes directly to customers. These relational settings focus on improving the fulfillment strategies adopted by companies. An example is MIT policy. • Multilevel inventory coordination. This means of coordination across the supply chain helps to accelerate order processing, decrease inventory levels, and increase customer satisfaction. These strategies focus on the supply-planning phase. Companies try to coordinate their production schedules with the customers’ demand. These partnerships in the supply chain are based on information sharing, technology sharing, and process integration. They help decrease the uncertainty in planning, increase the level of control, and improve the efficiency of operations. However, forging into a relational setting is not very easy and creates many obstacles. Table 1 summarizes the advantages and drawbacks of partnerships. Another collaboration milestone is the choice of the transport provider. As Figure 2 shows, any firm has the option of selecting its goods-transport mode. It could be using its own fleet, the United States Postal Service (USPS), common carriers such as Ryder and FedEx, or a third-party logistics (3PL) provider. A firm’s choice among these alternatives depends on the nature of its operation and objectives. Companies interested in maintaining control over their distribution usually operate their own fleet. Others interested in fast delivery with higher flexibility to market changes are likely to adopt common carriers. The USPS, for example, possesses a distribution channel set for deliveries to homes and is likely to capture the highest market share relating to that sector. On the other hand, fast courier transporters (for example, FedEx and UPS) are set to deliver to business districts. Finally, firms lacking the internal resources to manage logistics operations resort to 3PLs. These 3PLs are broadening their scope of operation by introducing logistics solutions through which they offer to manage their customers’ supply chain (e.g., UPS Logistics). The company’s choice of a transport and logistics provider should depend on the provider’s reliability, relationships record, fleet size, and ability to adjust to market changes. Finally, firms usually engage in partial or full outsourcing of some of their operations because resources are not available internally, function is difficult to manage, risks are shared, service quality is improved, cost reductions are achieved, and advanced technologies become within reach.

TABLE 1

Electronic commerce is affecting freight movement by reducing the flow of certain materials and increasing the transport of other commodities. The effect of the Internet on the freight-transportation sector includes its impact on • Just-in-time patterns and trends, • Current and prospective North American Free Trade Agreement flows, • Intermodal freight movements, • Urban goods distribution, • Inland movement of goods in foreign trades, • Travel and tourism requirements, both intercity and international, and • Major new trade corridor flows (10). In the following paragraphs, general insights into the effect on the transport of goods are presented, differentiating between B2C and B2B sectors. First, possible reduction of the total amount of freight transportation caused by the B2C sector could be explained by • Electronic materialization of many goods; • Elimination of personal transportation because electronic shopping might displace an object that would have been shipped anyway; and • Elimination of business transportation in some cases, in which the retailer would choose the best distribution center to ship from. On the other hand, B2C-related transportation could increase because of • More freight being transported; • Probable inefficiency in distribution systems of e-tailers; and • Increase in international imports, which leads to an increased demand for transportation. Possible increase of the total amount of freight transportation caused by the B2B sector could be explained by • Increase in globalization and international trade among companies; • Move toward customer-driven production policies, which might increase the frequency of shipments to retailers in order to reduce inventory levels; and • Adoption of faster delivery modes such as trucks and airplanes.

Advantages and Drawbacks of Partnerships

No Partnership Duration Information Exchange Transferability

13

Relational Setting

Drawbacks

One time

Long term

Conflicts

Little

Multiple way sharing

Fear of information sharing

Easy

Very difficult

—

Focus

Profit focus

Supply chain focus

Each party wants the biggest portion

Planning and Goals

Short term

Long term

Conflicts

Risks

Individual

Shared

—

14

Paper No. 02- 2913

However, the increase in business-related transportation is likely to be offset by the decrease caused by • Increase in collaboration among supply-chain parties, which helps in bundling demands; • Increased access to customers’ demand information, which reduces unwanted inventory on retailers’ shelves; • Reduction in mistaken orders, which eliminates unneeded delivery trips; and • Increase in auctioning activity that helps in filling empty spaces on transportation modes, allowing for a more dynamic market. Thus, the net effect on freight activity should be carefully monitored. Finally, cost reductions achievable through ICT are a major incentive for companies to invest in new developments. In the following section, the authors quantify cost reductions achievable through information sharing when a VMI strategy is applied (presented in Figure 2) in a two-echelon distribution channel.

COST REDUCTIONS THROUGH VMI VMI is a logistics distribution strategy through which the supplier manages inventory at customers’ sites and decides on replenishment policies, subject to stipulated levels of availability and service. The supplier benefits by reducing its inventory levels, reducing customers’ demand variability, and improving routing strategies; customers benefit by reducing resources dedicated to manage inventories and by decreasing their stock-outs, thus increasing their revenues. VMI is made possible by installing technological equipment at participating sites to allow the supplier to keep track of customer demand and inventory. The decrease in the cost of applicable technologies, such as EDI and the Internet, has accelerated the adoption of this strategy across several industries. The objective of the supplier in VMI is to decide on distribution tactics that will minimize the inventory and transportation costs across the supply chain. Many researchers have investigated integrated inventory and transportation problems, and several models that capture these scenarios are found in the literature. Harris proposed the classical economic order quantity (EOQ) model (11). The model has been applied by Blumenfeld et al. (12) and Burns et al. (13). Bertazzi et al. further improved the proposed models by constraining the shipments to a given set of discrete frequencies (14); this eliminated some infeasible solutions obtained by previous models (from a practical point of view). However, since the 1980s, researchers have mainly approached integrated-inventory and vehicle-routing problems by solving the inventory-routing problem (IRP). The IRP solution determines which customers should be visited in a specified planning horizon and optimal routing solutions to the supplier’s fleet, which usually reduces the question to a vehicle-routing problem. Major research work started to emerge in the 1980s. Due to the complexity of the general IRP (longterm dynamic stochastic setting), almost all of the research work available in the literature has concentrated on solving the short-term version. Most of the early work focused on solving the short-term IRP covering a single day (15–17 ). Later, this problem was expanded to span several days (18–21). The main differences among the various approaches lie in the customer-selection procedures and in incorporation of long-term effects of short-term decisions. Other research work investigated the asymptotic behavior of proposed simple poli-


cies (22–25). Finally, Bertazzi et al. analyzed a distribution problem, in which a product must be replenished from a supplier to multiple retailers (26). They analyzed the variation in total costs across the supply chain when the objective function (corresponding to different decision makers) is altered. The following subsections briefly describe the problem analyzed by Rabah and Mahmassani (27 ) and summarize the results of the simulation experiments.

Description of Problem The main objective of the analysis is to quantify cost reductions achievable through information sharing in a two-echelon distribution system composed of one supplier and multiple retailers. To achieve this objective, two different scenarios are modeled: no information sharing (NIS) and information sharing (IS). Figure 3 describes the decision policies involved in both scenarios. In the NIS scenario, an EOQ model is adopted for the retailers’ ordering policies. Each retailer is charged a fixed transportation fee per delivery in addition to the ordering cost, both of which affect its order quantity. In the case of stock-outs due to the stochastic nature of the daily demand, retailers are responsible for the incurred losses. However, if retailers decide to share demand information with the supplier (IS scenario), the supplier becomes responsible for devising a delivery strategy to his or her customers (VMI strategy) and thus bears the consequences of any stock-outs at retailers’ sites. In this setting, the supplier develops a strategy at the beginning of the planning horizon and abides by it unless a stock-out occurs, in which case the supplier makes an urgent delivery the following day. In both scenarios, daily inventory-holding, transportation, and stock-out costs are monitored. Rabah and Mahmassani provide further understanding of the problem (27).

Simulation Experiments and Results Following are descriptions of the system features specified in the simulations, the principal factors that were varied across experiments, and performance descriptors considered in the analysis. Finally, simulation results based on randomly generated instances are analyzed. The logistics network considered in all the experiments consists of 50 retailers [n = 50, i ∈ I = (1, . . . , n)] served by a common supplier. Moreover, the logistics network is assumed to be symmetrical and complete, that is, the transportation cost between any two parties is randomly uniformly generated in the interval 10 to 1000. All retailers are assumed to observe an average daily demand µi in the interval 50 to 150 (uniformly distributed), with a base coefficient of variation (COVi) uniformly generated in the interval 0.1 to 0.3; the daily demand at retailer i on day t, rit, is a truncated normally distributed random variable with mean µi and standard variation σi = COVi × µI (negative and high positive values are eliminated). The inventory holding cost, hi, i ∈ I, is uniformly generated in the interval 0.5 to 1. The stock-out cost per unit of shortage is equal to $10 at retailers and the supplier; however, in all the generated instances, the supplier adopts a 99% level of service (z0.01 = 2.33) to eliminate stock-outs. Finally, all the experiments were simulated along a time horizon H = 30 days, whereas the lead-time Li, i ∈ I, is equal to 5 days. In addition to the IS versus NIS scenarios, the experimental factors under investigation in the simulations are (a) inventory holding cost at the supplier, h0; (b) COV of the daily demand; (c) level of ser-


Paper No. 02- 2913

15

Customer Demand - Planning horizon: T (days) - Normally distributed - IID across retailers and days

Retailers and Supplier Decisions

No Information Sharing (NIS)

Retailer Ordering Decisions - Specify customer

Supplier Delivery Decisions - Fixed delivery

level of service - (si, Si) policy according to fixed lead time (dictated by supplier), and fixed costs charged per delivery

Retailer Costs - Ordering, Inventory, Stock-out Cost - Transportation Cost

Information Sharing (IS)

Retailer Ordering Decisions

Supplier Delivery Decisions

schedule dictated by retailers’ demands - Orders a daily quantity S0 that eliminates stock-outs - Performs daily routing

- No direct control on replenishment schedule - Place order in case of stock-out - Specify minimum and maximum inventory levels

- Develops delivery strategy along the time horizon according to forecasted demand - Has perfect information about customers’demand - Abides by starting policy unless a stock-out occurs - Orders a daily quantity according to encountered demand - Performs daily routing

Supplier Cost - Inventory Cost

Retailer Costs - Inventory Cost

Supplier Costs - Inventory Cost - Stock-out Cost - Transportation Cost

FIGURE 3 Comparison of policies under NIS and IS scenarios (IID independently identically distributed).

vice (LOS) at retailers; and (d ) fixed ordering cost K incurred by a retailer. First, the inventory holding cost at the supplier is likely to affect the frequency of shipments to the retailers’ sites in the IS setting, and thus the cost components of the solution and the associated total cost are likely to be altered. Second, two intervals were considered for the retailers’ demand COV. An increase in the demand variability is likely to increase the inventory cost at retailers and supplier due to an increase in the safety stock that each party should maintain. Moreover, it is likely to affect the number of stock-outs that occur along the time horizon. Third, two values for the LOS adopted by retailers were used: 95% (z0.05 = 1.65, reference case), and 96% (z0.04 = 1.75). The higher level of service represents retailers with high-value items. These retailers usually keep high safety stocks to avoid stock-outs, which affect their revenues considerably. Fourth, two values for the retailers’ ordering cost are used: K = 10 (base case), and K = 1000. A high ordering cost affects the orderup-to level that retailers adopt in the NIS setting and thus decreases the frequency of orders placed by retailers. Four principal system descriptors are analyzed in the comparison between the NIS and IS scenarios. First, the change in the inventory cost at retailers is analyzed. In the first scenario, this cost is mainly

dependent upon the level of service that retailers adopt and the lead time assumed, whereas, in the second scenario, it is controlled by the supplier’s shipment frequency. Second, the change in the supplier’s inventory cost is investigated. Third, variations in the stock-out costs are explored. Stock-out costs are a major concern for companies in competitive market environments, in which a stock-out might result in the loss of a customer to other companies. IS is likely to improve customers’ level of service. Finally, variations in transportation costs are examined. These costs are likely to be affected by the number of trips performed by the supplier in both scenarios. All the presented costs were monitored daily along the planning horizon; however, total costs are reported. For each of the aforementioned scenarios, five data sets have been randomly generated (Table 2), amounting to a total of 60 simulated instances. Simulation experiments were performed using Fortran. Total cost reductions were achieved in all the generated instances. Improvements varied between 0.1% and 18.26% (Figure 4). The lowest decrease is noted under the high-demand-variability scenario (COV = 0.1 to 0.6), which is expected since the supplier is following its initial, a priori strategy, which assumes a constant daily demand at the retailers’ sites. The highest saving is achieved under

16

Paper No. 02- 2913


TABLE 2

Experimental Factors and Generated Instances

Scenario Base Case Low h0 High h0 High COV High LOS High K

h0 0.75 0.50 1.00 0.75 0.75 0.75

Parameter Values LOS COV 0.1-0.3 95% 0.1-0.3 95% 0.1-0.3 95% 0.1-0.6 95% 0.1-0.3 96% 0.1-0.3 95%

K 10 10 10 10 10 1000

the high-ordering-cost scenario (K = 1000). This is also expected since in the NIS setting ordering cost represented about 17% of the total cost, whereas this cost element is eliminated under the IS scenario. The scenario definitions follow Table 2. For example, the high-ordering-cost scenario is one where K = 1000 with all other parameter values at their base case level (h0 = 0.75, COV = 0.1 to 0.3, LOS = 95%). Although adopting a VMI strategy decreases the total cost across the supply chain, it is important to examine the benefit for each collaborating party. In the following paragraphs, the authors analyze the three major components of the total cost: retailers’ inventory cost, supplier’s inventory cost, and transportation cost. Figure 5 presents the variation in the retailers’ inventory cost when a VMI strategy is applied compared to the NIS scenario. Increase in this cost is observed across all instances. The percentage increase varied between 15.66% and 29.24%. This rise in cost is caused by a higher shipping frequency that the supplier adopts to reduce its own inventory cost and eliminate stock-outs at retailers’ sites. Moreover, this increase in the retailers’ inventory cost is justified by the assumption made in the IS scenario in which the maximum inventory level at retailer i is taken to be equal to his economic ordering quantity computed using the original lead time (Li = 5 days), thus keeping the inventory level high. To reduce this jump in cost, the supplier could assume a lower upper limit (e.g., using a 2-day lead time), thus decreasing the inventory cost at the retailers. However, this strategy is likely to increase the probability of stock-outs at its customers, thereby increasing its cost. A decision should be made according to the associated value of the product being shipped and to the variation in stock-out and inventory costs. Moreover, the lowest increase in retailers’ inventory cost is noted for retailers with high ordering charge. By adopting

Number of Generated Instances IS NIS 5 5 5 5 5 5 5 5 5 5 5 5

Total Instances 10 10 10 10 10 10

new technologies, these retailers can reduce these costs and speed up their processes while achieving higher accuracy levels. Finally, although this rise in cost is a disadvantage for retailers, it is compensated by the elimination of the ordering and stock-out costs. Another advantage of adopting a VMI strategy is that retailers can decrease their resources dedicated to manage inventories and thus further reduce their operating costs. The supplier’s decrease in inventory holding cost varied between 14.65% and 19.41% (Figure 6). Information sharing, in which the supplier directly accesses the retailers’ customers demand and plans its replenishment policy accordingly, primarily drives this decrease in cost. The IS process decreases the variability of the demand observed by the supplier, thus lowering its safety-stock levels. Moreover, the higher shipping frequency to retailers’ sites decreases the amount of inventory available at the supplier and thus decreases its holding cost. However, this reduction in cost is diminished by stock-outs and transportation costs for which it becomes responsible. Finally, another aspect that must be examined by the supplier is the technology costs that it should invest to be able to track the retailers’ inventory levels; however, these costs are decreasing and are likely to be overshadowed over the long run by VMI benefits. Finally, analysis of the change in the transportation cost shows that it has been reduced in all instances other than the one with high demand variability, under which it increased by 20.33% (Figure 7). The percentage decrease varied between 13.04% and 17.44% on all tested instances. However, when a VMI strategy was applied, the number of visits made by the supplier to retailers along the time horizon of interest at least doubled compared to the NIS setting (Table 3). The percentage increase in the number of visits fluctuated between 121.55% and 225.55%, corresponding to

20

18.26

Decrease (%)

16 12 8 3.89

4 0.27

1.78 0.10

1.53

0 h0=0.5

h0=0.75

h0=1.0

COV=0.1-0.6

LOS=96%

K=1000

Generated Instances FIGURE 4 Percentage decrease in total cost when shifting from noncollaborative to collaborative environment (h 0 0.75 corresponds to the base case).


Paper No. 02- 2913

35 30

28.93

29.24

26.64

26.45

Increase (%)

25

23.04

20 15.66

15 10 5 0 h0=0.5

h0= 0.75

h0= 1.0

COV=0.1-0.6

LOS=96%

K=1000

Generated Instances FIGURE 5 Percentage increase in retailers’ inventory holding cost when shifting from an NIS to an IS setting (h 0 0.75 corresponds to the base case).

25

Decrease (%)

20

18.42

18.33

19.41

19.20 16.60 14.65

15 10 5 0 h0=0.5

h0=0.75

h0=1.0

COV=0.1-0.6 LOS=96%

K=1000

Generated Instances FIGURE 6 Percentage variation in the supplier’s inventory holding cost when applying a VMI strategy (h 0 0.75 corresponds to the base case).

20 15

13.04

17.44

13.28

16.73

15.47

LOS=96%

K=1000

10 Change (%)

5 COV=0.1-0.6

0 -5

h0=0.5

h0=0.75

h0=1.0

-10 -15 -20 -20.33 -25 Generated Instances

FIGURE 7 Percentage change in the transportation cost when shifting from an NIS to an IS scenario (h 0 0.75 corresponds to the base case).

17

18

Paper No. 02- 2913


TABLE 3 Increase in Number of Visits Made to Retailers When Adopting a VMI Strategy (h 0 0.75 Corresponds to the Base Case)

Generated Instance Number of Visits Using NIS Number of Visits Using IS Percentage Change

h0=0.50

h0=0.75

h0=1.00

COV=0.1-0.6

LOS=96%

K=1000

207

207

207

200

206

185

459 121.55

475 129.47

506 144.35

650 225.55

459 123.03

459 147.36

the instances with lowest inventory cost at the supplier (h0 = 0.5) and high demand variability (COV = 0.1 to 0.6), respectively. In the first case, since the supplier is minimizing the total cost in the VMI setting, the increase in the supplier’s shipment frequency is the lowest among other cases because it has a lower inventory cost than the retailers. On the other hand (as mentioned earlier), when the demand variability is high, stock-outs that induce urgent deliveries occur at a higher frequency. This requires the supplier to deviate from its initial schedule, so that it experiences the highest percentage increase in the number of visits. Finally, although the number of visits more than doubled in the five other experiments, the transportation cost decreased, reflecting the advantage of a VMI strategy in which the supplier bundles its customers into efficient routes, which lowers its transportation cost.

CONCLUSION The framework presented in this paper has identified the principal actors, processes, and interconnections that are likely to determine the still unfolding impacts of ICTs on logistics and supply-chain operations. The framework reflects the transformation of the traditional supply chain into a virtually forged one and its link to the transportation sector. The analysis focused on the increase in collaboration among supply chain parties, the types of coordination in the supply chain, and the effect of ICTs, especially the Internet, on freight movement. To illustrate and quantify cost savings that might result from information sharing among supply-chain participants, the paper considered the example of a VMI strategy in a two-echelon distribution channel. Reductions in total cost were achieved across all simulated instances. Moreover, results showed that the supplier benefits by reducing its inventory cost due to the decrease in demand variability that it observes, which allows it in turn to decrease its safety stock. The transportation cost decreased in almost all cases, although the number of visits made by the supplier increased under the IS scenario. This corresponds to another advantage of VMI, namely, the ability of the supplier to bundle its demands and thus reduce the total transportation cost. However, inventory cost at the retailers’ sites increased, in accord with the assumption made in this analysis that the retailer maintains an inventory capacity corresponding to the original lead time (with NIS). Decreasing this capacity is likely to decrease the retailers’ inventory cost but increase the stock-outs, which were eliminated in the simulations presented in the paper. Analyzing the tradeoff between the decrease in inventory cost and increase in stock-out cost represents a direction of ongoing and future work by the authors. Other future research directions include expansion of the supply chain into a multiechelon distribution channel.

ACKNOWLEDGMENTS This paper is based on research supported by the Southwest University Transportation Center and the Advanced Technology Program of the Texas Higher Education Coordinating Board. The work has benefited from comments and interaction with Patrick Jaillet and Miguel Figliozzi. The authors are grateful to Maria Grazia Speranza and Luigi Bertucci of Brezia University, Italy, for supplying the code used in this work to analyze the IS VMI scenario. The authors remain solely responsible for the content of the paper.

REFERENCES 1. Council of Logistics Management. CLM’s Definition of Logistics, 2001. www.clm1.org. 2. Metz, P. Newsletters, 1998. MIT Center for Transportation Studies. www.mit.edu. 3. Mahmassani, H. S. ICT and the Logistics Process: Third Party Providers and Freight Outsourcing. Impact of Information and Communication Technology (ICT) Seminar. Transport Canada, 1998. 4. D’Amours, S., B. Montreuil, P. Lefrancois, and S. Francois. Networked Manufacturing: The Impact of Information Sharing. International Journal of Production Economics, Vol. 58, 1999, pp. 63–79. 5. Wang, H. A Conceptual Model for Virtual Markets. Information & Management, Vol. 32, 1997, pp. 147–161. 6. Lancioni, R., M. Smith, and T. Oliva. The Role of the Internet in Supply Chain Management. Industrial Marketing Management, Vol. 29, 2000, pp. 45–56. 7. Delaney, R. V. State of Logistics Report, 1998. Cass Press Conference 1999. www.cassinfo.com. 8. Simchi-Levi, D., P. Kaminski, and E. Simchi-Levi. Introduction to Supply Chain Management. In Designing and Managing the Supply Chain. McGraw-Hill, New York, 2000, pp. 1–13. 9. Simchi-Levi, D., P. Kaminski, and E. Simchi-Levi. Information Technology for Supply Chain Management. In Designing and Managing the Supply Chain, McGraw-Hill, New York, 2000, pp. 215–248. 10. Romm, J., A. Rosenfeld, and S. Hermann. The Internet Economy and Global Warming, Cool Companies, 1999. www.cool-companies.org. 11. Harris, F. W. How Many Parts to Make at Once. Factory Magazine of Management, Vol. 10, 1913, pp. 135–136. (Reprinted in Operations Research, 1990, pp. 947–950.) 12. Blumenfeld, D. E., L. D. Burns, J. D. Diltz, and C. F. Daganzo. Analyzing Trade-Offs Between Transportation Inventory and Production Costs on Freight Networks. Transportation Research, Vol. 19B, 1985, pp. 361–380. 13. Burns, L. D., R. W. Hall, D. E. Blumenfeld, and C. F. Daganzo. Distribution Strategies That Minimize Transportation and Inventory Costs. Operations Research, Vol. 33, 1985, pp. 469– 490. 14. Bertazzi, L., M. G. Speranza, and W. Ukovich. Minimization of Logistic Costs with Given Frequencies. Transportation Research, Vol. 31B, 1997, pp. 327–340. 15. Federgruen, A., and P. Zipkin. A Combined Vehicle Routing and Inventory Allocation Problem. Operations Research, Vol. 35, No. 5, 1984, pp. 1019–1037.


16. Golden, B., A. Assad, and R. Dahl, R. Analysis of a Large-Scale Vehicle Routing Problem with an Inventory Component. Large-Scale Systems, Vol. 7, 1984, pp. 181–190. 17. Chien, T. W., A. Balakrishnan, and R. T. Wong. An Integrated Inventory Allocation and Vehicle Routing Problem. Transportation Science 23, 1989, pp. 67–76. 18. Dror, M., M. Ball, and B. Golden. A Computational Comparison of Algorithms for the Inventory Routing Problem. Annals of Operations Research, Vol. 4, 1985, pp. 3–23. 19. Dror, M., and M. Ball. Inventory/Routing: Reduction from an Annual to a Short-Period Problem. Naval Research Logistics, Vol. 34, 1987, pp. 891–905. 20. Trudeau, P., and M. Dror. Stochastic Inventory Routing: Route Design with Stockouts and Route Failures. Transportation Science, Vol. 26, No. 3, 1992, pp. 171–184. 21. Huang, L., J. F. Bard, P. Jaillet, and M. Dror. Inventory Routing Problem with Satellite Facilities. Technical Report. University of Texas at Austin, 1997. 22. Anily, S., and A. Federgruen. One Warehouse Multiple Retailer Systems with Vehicle Routing Costs. Management Science, Vol. 36, No. 1, 1990, pp. 92–114.

Paper No. 02- 2913

19

23. Gallego, G., and D. Simchi-Levi. On the Effectiveness of Direct Shipping Strategy for the One-Warehouse Multi Retailer R-Systems. Management Science, Vol. 36, No. 2, 1990, pp. 240–243. 24. Anily, S., and A. Federgruen. Two-Echelon Distribution Systems with Vehicle Routing Costs and Central Inventories. Operations Research, Vol. 41, 1993, pp. 37–47. 25. Bramel, J., and D. Simchi-Levi. A Location Based Heuristic for General Routing Problems. Operations Research, Vol. 43, No. 4, 1995, pp. 649–660. 26. Bertazzi, L., G. Paletta, and M. G. Speranza. Deterministic Orderup-to-Level Strategies for the Minimization of the Logistic Costs in Distribution Systems. In New Trends in Distribution Logistics (M. G. Speranza and P. Stahly, eds.), Springer-Verlag, New York, 1999, pp. 179–196. 27. Rabah, M. Y., and H. S. Mahmassani. Impact of Electronic Commerce on Logistics Operations: A Focus on Vendor Managed Inventory (VMI) Strategies. Research Report SWUTC/02/167227-1. Southwest Region Transportation Center, University of Texas at Austin, 2002. Publication of this paper sponsored by Committee on Freight Transportation Planning and Logistics.