who in each network. This is because an important factor in choosing between incentive tools is to be informed on the alternatives that other players have and the costs associated with them. Therefore, the type of players and their interrelations are likely to determine the type of incentive tools employed.
Table3.7:CaseStudiesonInventoryManagement
ReferenceProduct(in)RecoveryProduct(out)SenderCollectorProcessorFutureOrganizerReturnDriversoptioncustomerreason[59]ind.goodsre-useind.goodsinternalcustomerCERNinternalCERNreiumbursement,economics(Europe)customers(bring-back)customersend-of-use[59]cons.goodsre-salecons.goodscustomersmailordermailordercustomersmailorderreimbursementeconomics,(Europe)companycompany(samemar-ket) companylegislation [59]sparepartsrepair,sparepartsserviceserviceRefineryinternalserviceeconomics(Europe)(refinary)re-useengineersengineerscustomer[76]distribut.re-distribut.(product-in)retailersCoca-ColaCoca-ColaoriginalCocaColafunctionaleconomics(N.A.)itemschainchain[79]sparepartsrepairsparepartsCaracasCaracasCaracasCaracasCaracasserviceeconomics(S.A.)(railways)subwaysubwaysubwaysubwaysubway[82]sparepartsrepairsparepartstelephoneLucentLucentsameLucentserviceeconomics(Europe)(telephones)companiesTechnologiesNederland TechnologiesNederland chainTechnologiesNederland[101]ind.goodsrepair,ind.goodsbusinessIBMend-of-lifeeconomics(Europe)(IBM)refurbish.spare-partscustomersetc.[154]cons.goodsremanufact.,(product-in)customersspecializedspecializedmanufacturersend-of-lifeeconomics(Europe)(powertools) recyclingothermater.partyfacility(pro-active)
[192]ind.goodsrepair,ind.goodsUKUKUKUKUKserviceeconomics(Europe)(aircraften-gine) refurbishingAirForceAirForceAirForceAirForceAirForce [222]cons.goodsreuse,(product-in)usersNorwegianTACorsameNorwegianend-of-useeconomics,(aidequip.)refurbish.,TechnicalAid specializedmarketNationalservicecorporate (Europe)retrieval,...Centers(TAC) playerInsuranceAdministra-tion citizenship [223]cons.goodsre-salecons.goodscustomer3rdpartyWehkampsameWehkampreimbursementeconomics(Europe)log.provider(catalogue)market(legislation)[242]distribut.items re-distribut.distribut.items businessBavariaBavariaBavariafunctionaleconomics
(Europe)(beerkegs)customersagentsBavaria[256]cons.goodsremanufact.cons.goodsconsumerphotoKodaksameKodakserviceeconomics(single-useshops,chain(Europe,photootherN.A.)cameras)retailers[266]sparepartsremanufact.sparepartsImporterVolkswagenVolkswagenImporterVolkswagenserviceeconomics(Europe)(cars)organizations(inKassel)(inKassel)organizations(inKassel)
Commercial returns (3 cases)
Commercial returns occur in a B2B or in a B2C setting (see Section 3.2), where the buyer has a right to return the product, usually within a certain period. In the B2B setting, there are all kind of contracts lowering the risk of e.g. the retailer by giving him the opportunity to return what is not sold.
In this situations, returns are likely to be in bulk and at the end of a season.
In the B2C setting, consumers are given, or they have by law, the right to return a product if the product does not really meets his/her expectations.
Sanders et al. [223] describe how the inventories of products are controlled within Wehkamp, a Dutch mail order company, selling all kinds of consumer goods to the Dutch and Belgium market. Two types of products are dis-tinguished: seasonal and non-seasonal products. For the first, the company employs an amended version of the newsboy model taking into account re-turns. For the latter, the inventory management is done according to an (R, S) policy.
De Brito and Dekker [59] investigate the distribution of the return lag, i.e.
the time between the purchase and the return of an item, and its consequences for inventory management (see Chapter 6). Two cases of commercial returns are considered, viz. a mail order company and and the warehouse at the center for nuclear research, CERN.
Service return (6 cases)
Within service systems like repair systems returns occur basically in two ways.
First of all, the products themselves may be brought or sent to a center for repair. If the repair is successful, they are brought back, else, a new product or system needs to be bought and the one that failed is discarded. Secondly, if one needs a continuous functioning of the product or system, one may directly restore functionality by replacing a spare part. The failed part is then repaired, after which it will enter the inventory of spare parts. The cases found are described below in detail.
D´iaz and Fu [79] study a 2-echelon repairable item inventory model with limited repair capacity. For several classes of arrival processes they develop an analytic expression for the number of items in queue at the different stages of the system. They analyze the impact of the capacity limitation and compare the performance of their approach with an uncapacitated METRIC type of model. Both models are applied to the case of spare parts management at the Caracas subway system.
Donker and Van der Ploeg [82] describe how the optimal stock of reparable service parts of telephone exchanges is determined within Lucent Technolo-gies Netherlands. They use an amended METRIC model, where the service
measure is fill rate (i.e. the percentage of demand that can immediately be fulfilled from stock) and there is no budget restriction for service parts.
Moffat [192] provides a brief summary of a Markov chain model for ana-lyzing the performance of repair and maintenance policies of aircraft engines at the Royal Air force.
Van der Laan [266] describes the remanufacturing chain of engines and automotive parts for Volkswagen. It is somewhat similar to the engine re-manufacturing case with Mercedes Benz in the previous section (see Driesch et al. [85]).
A special example of service returns is the one described by Toktay et al. [256] on Kodak’s single use camera. Customers return it, so they can develop the film. Printed circuit boards for the production of these cameras are either bought from overseas suppliers or remanufactured from the cameras returned by the customers via photo laboratories. The issue is to determine a cost-efficient order policy for the external supplies. Major difficulties arise from the fact that return probabilities and market sojourn-time distribution are largely unknown and difficult to observe. The authors propose a closed queuing network model to address these issues. They assess the importance of information on the returns for the control of the network.
De Brito and Dekker [59] investigate the distribution of the return lag, i.e.
the time between the purchase and the return of an item for a spare parts warehouse at a petrochemical plant.
End-of-use returns (1 case)
This return reason concerns items that are only temporarily with the user.
The product may e.g. be leased or rented.
Rudi et al. [222] discuss the product recovery actions of the Norwegian na-tional insurance administration. This public entity retrieves no longer needed wheel chairs, hearing aids and similar products provided to people with hand-icaps. They assess how many are needed to meet all demands.
End-of-life returns (2 cases)
Fleischmann [101] describes the dismantling of computers at the end of their life-cycle into useable spare parts with IBM. This source nicely combines with return obligations and it is a cheap source for spare parts for systems on which one does not want to spend too much. The problems identified were a lack of knowledge of the content of the computers, as well as, an insufficient information system to handle the operations.
Klausner and Hendrickson [154] develop a model to determine the optimal buy-back amount to guarantee a continuous flow to remanufacturing
power-tools. The authors apply the model to the actual voluntary take-back program in Germany.
3.6.2 Preliminary Insights
We have grouped the presentation of the cases according to the why-returning typology, i.e. according to the return reason as follows:
• Functional returns,
• Commercial returns,
• Service returns,
• End-of-use,
• End-of-life
This is because this seems a natural way of grouping and discriminating the reverse logistics issues raising from each inventory system. Other authors have done roughly the same (see Dekker and van der Laan, 2003).
Many have defended that product data are essential for efficient handling of returns. For instance, Kokkinaki et al. (2004) provide an example of the value of information for disassembly. Other authors have investigated the impact of data, with respect to returns, on inventory management perfor-mance (Kelle and Silver, 1989 ; Toktay et al., 2000; De Brito and Van der Laan, 2003). Yet, there is still room to model the impact of having a priori information on what can be recovered, i.e. on which parts are likely to be recoverable. In practice, techniques to forecast return behavior would have to be enriched with broader explanatory variables. We refer to Toktay (2003) for a discussion of factors influencing returns, which are potential explanatory variables in advanced forecasting models.
Concluding, it remains to be investigated in which degree inventory sys-tems’ characteristics like timing, quality and degree of control, are determined by the return reason why-returning, as well as on the type of product (what ).