Optimal radio frequency identification deployment in a supply chain network.docx

1、Optimal radio frequency identification deployment in a supply chain networkOptimal radio frequency identification deployment in a supply chain network Seokcheol Changa, , Diego Klabjana, and Thomas Vossenb, , a Industrial Engineering and Management Sciences, Northwestern University, Evanston, IL, US

2、Ab Leeds School of Business, University of Colorado, Boulder, CO 80309-0419, USAReceived 7 July 2009; accepted 8 January 2010. Available online 29 January 2010. AbstractIn recent years, the use of radio frequency identification (RFID) within supply chain management has received considerable attentio

3、n. A significant upfront investment together with uncertain return on investment, however, remain impediments to deployments of RFID systems. This paper proposes a novel approach to analyze the potential benefits of RFID systems. We suggest deployment strategies that determine an optimal location of

4、 RFID within a supply chain network. The resulting models can be used to evaluate the value of information with respect to item losses. In addition, we discuss methods for solving the resulting large-scale optimization problems, and analyze how various system characteristics impact optimal deploymen

5、t strategies as well as the overall RFID benefits.Keywords: RFID; Supply chain management; Facility locationArticle Outline1. Introduction 2. Literature review 3. RFID placement in serial networks 3.1. Lead time expressions 3.1.1. No recovery 3.1.2. Full recovery 3.1.2.1. Constant search time 3.1.2.

6、2. Proportional search time3.1.3. Partial recovery 3.1.3.1. Constant search time 3.1.3.2. Proportional search time3.2. Optimal placement4. RFID placement in general networks 5. Computational study 5.1. Experimental setup 5.2. Impact of loss probability 5.3. Impact of recovery probability 5.4. Impact

7、 of installation cost 5.5. Overlap probability of DCs in the path6. Summary and conclusions Acknowledgements Appendix A. Nomenclature References1. IntroductionOver the last few years, radio frequency identification (RFID) has emerged as an important new technology to track the movement of goods in a

8、 supply chain. An RFID system consists of three principal components. The first component is a tag with an electronic product code (EPC) to identify pallets, cases, or individual product items. In addition, an RFID system also requires readers that can initialize and receive information from the tag

9、s. The communication between the two is by radio waves. The final component is an information system with a data warehouse, which is used to store and manage the data captured from the readers. To operate an RFID system, readers send radio-wave signals over certain pre-specified frequencies. These s

10、ignals are received by tags in the vicinity of the readers. The tags then transmit stored data (such as the EPC number) back to the reader. The reader, in turn, decodes and transfers this information to the data warehouse together with the time and location stamp. As a result, RFID has several impor

11、tant advantages over traditional bar coding: items can be read from a distance without optical line of sight, multiple tags can be read simultaneously, and item specific data can be written on a tag (Gale et al., 2005). The RFID technology yields several important benefits within supply chains. Amon

12、g others, automated data tracking without human intervention offers faster processing, more accurate inventory records, and advanced shipping notices. Thus, product shrinkage, transaction errors, chargebacks, misplaced products, and incorrect product identification can all be reduced (Kang and Gersh

13、win, 2005 and zelkan et al., 2006). Moreover, unprecedented visibility of the product flows (see Delen et al., 2007, for examples of the detailed level of information that RFID can provide) provides consumers with the right amount of product at the right time while reducing safety stock and the risk

14、 of stockouts.As such, the use of RFID has enormous potential to increase supply chain efficiency. While companies in the U.S. invested $1 trillion on upgrading and improving supply chains in year 2000 alone (Gale et al., 2005) inefficiencies remain high. According to the National Retail Security Su

15、rvey, for example, the product shrinkage cost for U.S. retailers was over $31 billion in 2001; a similar number, $27.5 billion, has also been reported as the cost of shrinkage in Western Europe (Bamfield, 2004). In addition to this direct financial impact, however, this lack of visibility also has b

16、roader impact on supply chain performance, leading to excess inventory and/or lost sales. DeHoratius and Raman (2008) and Kang and Gershwin (2005) present empirical studies, which show that inventory record inaccuracies are substantial and widespread. As a result, companies may fail to order when ne

17、eded or place orders that are unnecessary. Thus, it is perhaps not surprising that investments in RFID technology have been growing rapidly, spurredin partby mandates from industry giants such as Wal-Mart and the U.S. Department of Defense. According to the market research group IDTechEx (2008), the

18、 global RFID market will reach $5.3 billion by 2008, with a fivefold increase expected by 2018.Nevertheless, the introduction of RFID systems poses a number of important issues. Impediments to deployment, for example, arise due to the absence of international standards, security and privacy concerns

19、, and lower read accuracy in environments with liquids or metals. In particular, the high costs of implementation often restrict the adoption of RFID systems in supply chain management, especially in light of the uncertain return on investment. Only about 30% of Wal-Marts top 100 suppliers, for one,

20、 have introduced full-scale RFID deployments and 95% of them did not expect return on investment within two years (Gale et al., 2005 and Chuang and Shaw, 2005).As a result, the justification of RFID investments remains a major challenge. Given that it is difficult to measure supply chain benefits th

21、at result from visibility improvements, practitioners often resort to rough estimates and superficial analysis (see Lee and zer, 2007, for an excellent discussion). Generally speaking, assessments of RFIDs return on investment proceed by comparing a “no-visibility” scenario, corresponding to the cur

22、rent situation, with a “full-visibility” scenario that is based on a complete deployment of RFID. These, however, can be viewed as extremes within a broad spectrum of “partial-visibility” options available to a company when deciding upon the extent and scope of their RFID investments. The main objec

23、tive of our work, therefore, is to analyze how different RFID deployment levels can benefit the supply chain. In doing so, our goal is to provide new insights on how visibility impacts supply chain performance, which may prove useful when companies contemplate RFID investments.To address these issue

24、s, we evaluate the placement of RFID deployments, i.e., the locations of RFID deployments in the supply chain network. A deployment presents a significant cost, which is typically estimated at up to several hundred thousand dollars at a single location (Simchi-Levi et al., 2007). While the cost of a

25、 reader itself is in the thousands of U.S. dollars, the majority of this amount results from software and integration needs. Decision makers have to consider deployment strategies that balance the benefits of RFID with their deployment costs at an aggregate network level. While recent research consi

26、ders the performance improvements that may result from RFID installations at a single location (zelkan et al., 2006 and Sounderpandian et al., 2006), we believe that considerations of these macro-level trade-offs have not been addressed before.We study approaches to evaluate such aggregate benefits

27、of RFID in a supply chain. Specifically, we propose models to determine the optimal locations of RFID systems in the network by trading off their potential benefits with RFID installation costs. To evaluate the potential benefits of an installation strategy, we consider its impact on the effective l

28、ead times in the system, which take into account the time spent to recover and/or reship items that are lost. Here, the effective lead times represent not only the time needed to receive a direct order from the supplier, but the order-to-delivery time that includes all of the steps between hitting t

29、he reorder point and placing a replenishment order as well as the time to receive goods. Lead times are an important factor in determining supply chain performance, given their impact on pipeline inventory levels, safety stock, transportation costs, etc. We believe that the cost benefits resulting f

30、rom RFID-induced lead time reductions provide an attractive and novel way of measuring the benefits of increased visibility.As such, our study concentrates on the marginal benefits that arise from an early detection of supply and shipment disruptions. Within this framework, we consider a variety of

31、model alternatives to address different system characteristics and recovery strategies, each resulting in a complex non-linear location problem. To solve these problems, we propose a novel solution methodology that applies dynamic programming techniques within a branch-and-price framework, together

32、with a heuristic transformation technique to obtain integral solutions. The contributions of our paper are therefore twofold. To the best of our knowledge, our work is the first one to evaluate the aggregate supply chain benefits of an RFID deployment at the system level. In addition, we also propose novel methodologies to solve the models that arise from our approach.Our models and solution methodologies do not explicitly use any particular features of RFID. Any technology that improves visibility through an ability to track goods can be deployed based on the models provided. R