Issues in reverse supply chains, part I: end‐of‐life product recovery and inventory management – an overview

Abstract

Recovery of used products has become a field of rapidly growing importance in reverse supply chain management. Product recovery includes collection, inspection/separation, disassembly, reconditioning/reuse, remanufacturing, and recycling. In real time situations, the collection of end‐of‐life (EoL) products from the customer and their return to the manufacturer is tedious and time consuming. Reduced product life cycles have increased the rate of product returns and disposals. Owing to shortened product economic life cycles, the recovery of value from EoL products is becoming a necessity. Companies have realised the value that they could recover by remanufacturing or recycling EoL products. Researchers have developed various models for product recovery network design, optimal inventory, production planning and control, remanufacturing, recycling, disposal, etc. The main purpose of this paper is to review the literature on EoL product recovery and inventory management issues in reverse supply chains and to outline some future directions for research on these issues.

Keywords:

• product recovery
• inventory management
• waste management
• supply chain management

1. Introduction

To respond to increasing competition and the decreasing life cycle of products, more firms are now embracing the concept of Supply Chain Management (SCM) to mange their business effectively and to increase their overall profitability. The tool to streamline and optimise the task of collecting end‐of‐life (EoL) products, refurbish them and sell them is referred to as Reverse Supply Chain Management (RSCM). Owing to the revolution in green manufacturing for the global market, reverse logistics concepts have become an important issue that can play a pivotal role in a company's competitive advantage and help strategic decision‐making. Srivastava (2007) classified the Green Supply Chain Management (GrSCM) literature into three broad categories: literature highlighting the importance of GrSCM; literature on green design; and literature on green operations. Carter and Ellram (1998) emphasised more on the environmental aspect of reverse logistics which they defined as the ‘process whereby companies could become more environmentally efficient through recycling, reusing, and reducing the amount of materials used’. Life Cycle Analysis (LCA) is a process for assessing and evaluating the environmental and resource consequences of a product through all phases (extracting and processing the raw material, production, distribution, use, remanufacturing, recycling and final disposal) of its life. Industries may use LCA to support product development so that overall environmental impact of the product can be minimised.

Product recovery management is a subset of reverse logistics, where all used and discarded products, components and materials are managed efficiently. Environmentally conscious manufacturing and product recovery has become a responsibility to Society and the environment, imposed primarily by government regulations and customer perspectives on environmental issues. Moyer and Gupta (1997) conducted a comprehensive survey of previous works related to environmentally conscious manufacturing practices, recycling, and the complexities of disassembly in the electronics industry. Gungor and Gupta (1999) presented the development of research in environmentally conscious manufacturing and product recovery (ECMPRO) and provided a state‐of‐the‐art survey of the published work in that area.

Product recovery aims to minimise the amount of waste sent to landfills by recovering the materials and parts from old or outdated products by means of reuse, recycling and remanufacturing. Thierry et al. (1995) presented an integrated supply chain framework to demonstrate reverse flows and recovery options such as repair, refurbishing, remanufacturing, recycling, etc. Figure 1 shows the framework of reverse supply chain activities. Recycling seeks to recover the material content of returned products by performing the necessary disassembly, sorting and reprocessing operations. On the other hand, remanufacturing preserves the product's identity and performs the required disassembly, sorting, refurbishing and assembly operations in order to bring the products to a desired level of quality. Guide et al. (1999) discussed the characteristics of the remanufacturing environment to distinguish it from other manufacturing environments and examined the production planning and control function of remanufacturing firms. A recent survey of disassembly was produced by Lambert (2003) and Williams (2006) summarised research to improve the demanufacturing processes through disassembly, automation, bulk recycling of ferrous and nonferrous metals, and plastics identification and separation.

A second key area in reverse logistics is inventory management. Proper control methods are required to integrate the return flow of EoL products into the manufacturer's material planning. Fleischmann et al. (1997) presented an overview on quantitative models for recovery production planning and inventory control. In their work, the authors discussed the recently emerged field of reverse logistics and subdivided the field into three main areas, namely: distribution planning; inventory control; and production planning. The authors not only emphasised inventory control of returned flows and reverse distribution, but also highlighted the lack of a general framework and mathematical model to support the reverse logistics environment. Guide et al. (2000) reviewed the literature on inventory systems with returns and Guide (2000) identified and described seven complicating characteristics of production planning and control activities for remanufacturing firms, such as the uncertain timing and quantity of returns; the need to balance returns with demands; the disassembly of returned products; the uncertainty in materials recovered from returned items; the requirement for a reverse logistics network; the complication of material matching restrictions; and the problems of stochastic routings for materials for remanufacturing operations and highly variable processing times. Minner (2003) reviewed inventory models with multiple supply options and discussed their contribution to supply chain management.

Figure 1 A framework for reverse supply chain activities.

Objectives and format

This paper is the first of three reviews by the authors that consider issues in reverse supply chain management. Subsequent papers, to be published in forthcoming issues of the Journal, will consider reverse distribution and the analysis of reverse supply chains. In addition to environmental regulation, product recovery and inventory management issues; problems related to waste management (finding better ways for waste treatment) are addressed. The specific objectives of this paper are: (1) to suggest a classification of available literature in the field of product recovery and inventory management issues; (2) to identify critical issues for each classification; (3) to identify emerging trends in the field of product recovery; (4) to suggest directions for future researchers in this field; and (5) as far as possible, to consolidate the available literature on product recovery and inventory management issues. The paper is organised as follows: after a brief introduction, the details of the research methodology are presented in Section 2. In Section 3, the existing literature has been classified based on product recovery and inventory management issues. Detailed discussion and the identification of critical issues are undertaken in Section 4. Finally, Section 5 closes the paper by offering conclusions and some suggestions for future research.

2. Research methodology

A literature survey was employed as the research methodology for this study to develop a framework (see Figure 1) for product recovery and inventory management issues in RSCM. The literature on RSCM was collected from journals covering operations management, supply chain, product recovery, operations research, environmental engineering, and information systems. The journals were published by various publishers, but especially Elsevier, Emerald, InderScience, Springer link, and Taylor and Francis. As we stated, our aim was to analyse those articles that directly related to product recovery and inventory management for RSCM.

3. Classification based on product recovery and inventory management issues

The literature available on RSCM is reviewed with a focus on product recovery and inventory management issues. The literature on product recovery and inventory management issues was classified into five categories: (1) environmental implications; (2) product recovery, with sub divisions to cover direct use, recycling, remanufacturing and repair; (3) common issues in product recovery, to include collection, inspection/separation and disassembly/scheduling; (4) inventory management, to include deterministic and stochastic models with the latter sub divided into periodic review and continuous review policy; and (5) waste management, to include disposal and pollution.

4. Critical issues

In this section, the available RSCM literature is reviewed and the issues are discussed based on the classification scheme presented in the previous section.

4.1 Environmental implications

Owing to growing environmental concern RSCM has received growing attention. The environmental impact associated with the reverse chain is examined with the help of a Life Cycle Analysis (LCA) study. Manufacturers will face completely new challenges when their products are subject to environmentally‐driven legislation. Environmental legislation and customer expectations increasingly force manufacturers to take back their products after use. Bloemhof‐Ruwaard et al. (1995) elaborated on the possibilities of incorporating green issues when analysing industrial supply chains and more generally on the value of using Operations Research (OR) models and techniques in Green supply chain management research. Fullerton and Wu (1998) reported that price‐based policies constitute a less challenging option in terms of implementation and monitoring. Examples of such policies included taxes on the use of virgin materials, recycling subsidies, disposal fees and deposit–refund requirements. Beamon (1999) investigated the environmental factors leading to the development of an extended environmental supply chain and described the additional challenges presented by the extension. Van Hoek (1999) presented a categorisation of green approaches and suggested the value‐seeking approach as the most relevant in greening the supply chain. Zhu and Sarkis (2004) examined the relationships between green supply chain management (GSCM) practice and environmental and economic performance. They evaluated the general relationships between specific GSCM practices and performance using moderated hierarchical regression analysis.

Georgiadis and Vlachos (2004) examined the impact of environmental issues on long‐term behaviour of a single product supply chain with product recovery. The behaviour of the system was analysed through a dynamic simulation model based on the principles of the system dynamics (SD) methodology. Sheu et al. (2005) formulated a linear multiobjective programming model that systematically optimised the operations of both integrated logistics and corresponding used‐product reverse logistics in a given green‐supply chain. Factors such as the used‐product return ratio and corresponding subsidies from a governmental organisation for reverse logistics were considered in the model formulation.

The European Union's End‐of‐Life Vehicle (ELV) Directive, which came into force in September 2000, aims to increase recovery of ELVs in order to reduce waste from EoL vehicles and improve environmental performance. It was designed to promote collection, reuse and recycling of these vehicles (Gerrard and Kandlikar 2005). Vlachos et al. (2007) tackled the development of efficient capacity planning policies for remanufacturing facilities in reverse supply chains, taking into account not only economic but also environmental issues, such as the take‐back obligation imposed by legislation and the ‘green image’ effect on customer demand. The behaviour of the generic system under study was analysed through a simulation model based on the principles of the system dynamics methodology. Sheu (2008) formulated a linear multiobjective optimisation model to optimise the operations of both nuclear power generation and the corresponding induced waste resource logistics.

4.2 Product recovery

Product recovery refers to the set of activities designed to reclaim value from a product at the end of its useful life. Product recovery is carried out mainly for three reasons: (1) governmental regulations; (2) market requirements; and (3) hidden economic value of solid waste. Pugh (1993) used mathematical models in evaluating resource recovery options. Johnson and Wang (1995) defined product recovery as a combination of remanufacturing, reuse, and recycling. Thierry et al. (1995) divided the product recovery process into repair, refurbish, remanufacture, cannibalise and recycle. Kannan et al. (2008a) proposed a multicriteria decision‐making model for selecting the collecting centre location in the reverse logistics supply chain model using the analytical hierarchy process and fuzzy analytical hierarchy process. Krikke et al. (1998) proposed a model for evaluating recovery strategies for the product without violating the physical and economical feasibility constraints. Melissen and de Ron (1999) defined recovery practices and provided relevant definitions and terminology.

4.2.1 Direct reuse

Some items may be reused directly (without prior repair operations) after cleaning and minor maintenance, typically they are used to contain other goods. Examples are reusable packages such as bottles (Torre et al. 2004), pallets or containers (Kroon and Vrijens 1995, Kelle and Silver 1989). Linton and Johnson (2000) described a decision support system for reuse and remanufacturing telecommunications equipment.

González‐Torre et al. (2004) focused especially on the joint implementation of environmental practices in collaboration with suppliers and customers. They analysed the differences existing in the relations between bottling/packaging firms that belonged to the food and drinks sector and their suppliers (fundamentally bottle/jar manufacturers) and customers (end consumers of the packaged or bottled products) in two European countries (Spain and Belgium) with different characteristics. Zikopoulos and Tagaras (2007) investigated the impact of uncertainty in the quality (refurbishability) of used product returns on the profitability of reuse activities in a single‐period system with one refurbishing site and two supplying collection sites facing stochastic demand.

4.2.2 Recycling

In the case of recycling, products are processed in order to obtain the desired quality following which they are reused. The goal of recycling is to recover the material without concern for the conservation of product structures. Examples for recycling are plastic recycling (Pohlen and Farris 1992), paper recycling (Pati et al. 2008), glass recycling (González‐Torre et al. 2004, González‐Torre and Adenso‐Díaz 2006) metal recycling from scrap (Hoyle 1995, Sigman 1995, Legarth 1996, Ayres 1997, Spengler et al. 1997, Johnson 1998, Khoei et al. 2002, Tsoulfas et al. 2002, Logožar et al. 2006), fibre optic cable recycling (Wright et al. 2005), sand recycling (Barros et al. 1998, Listes and Dekker 2005), electronic product recycling (de Ron and Penev 1995, Ashayeri et al. 1996; de Fazio et al. 1997, Nagel and Meyer 1999, Krikke et al. 1999a, Nagurney and Toyasaki 2005), electrical appliances recycling (Shih 2001), carpet recycling (Ammons et al. 1999, Louwers et al. 1999, Realff et al. 1999, 2004, Biehl et al. 2007), and automobile recycling (Isaacs and Gupta 1997, Bellmann and Khare 1999, 2000).

Carlson (1985) used weighted non‐linear goal programming to discuss the economic impacts of material recycling on energy recovery facilities. Pohlen and Farris (1992) identified a number of fundamental functions, including collection, separation, transitional processing, delivery and integration, within a typical reverse logistics channel in a plastic recycling case. In a study by Legarth (1996), it was clearly shown that without an intensified focus on recycling, we could not hope to fulfill even the most modest ambitions for sustainability in the use of metal primary resources in the future. Spengler et al. (1997) discussed two cases, one for recycling building debris and one for the recycling of by‐products in the German steel industry. Ayres (1997) reported that environmental problems would be the principal driver in the mining and metallurgical processing industrial sector in coming decades. Johnson (1998) described the reverse logistics systems for ferrous scrap in 12 North American manufacturing plants; examined the role of purchasing and other functions in the reverse logistics system; and assessed the contribution made by various departments. The recycling network was developed without government regulations or incentives.

Barros et al. (1998) proposed a two‐level location model for a sand recycling problem and considered its optimisation using heuristic procedures. They formulated a mixed integer linear programming model to minimise the total cost of the network. In Hirsch et al. (1998), a logistics planning tool (LOCOMOTIVE toolbox) was available that could handle ecological aspects with the advent of legislative decisions to reduce environmental pollution and to recycle products. Krikke et al. (1999a) discussed a PC‐monitor recycling case as a part of a broader pilot project at Roteb (the municipal waste company of Rotterdam in the Netherlands). They applied a two‐step procedure for optimising a recovery strategy for durable consumer products in a multi‐product situation. They determined product recovery and disposal strategy at the product level in the first step and, in the second step, a group recovery and disposal policy at the product group level. Kleineidam et al. (2000) described a modelling method which consisted of elementary models of standard production operations for production chains including recycling. The models were analysed using control theory.

Boon et al. (2002) investigated the critical factors influencing the profitability of EoL processing of PCs. They also suggested suitable policies for both PC manufacturers and legislators to ensure that there was a viable PC recycling infrastructure. Khoei et al. (2002) used the Taguchi method to optimise recycling processes and Hoyle (1995) used technical–economical constraints analysis for the case of aluminium recycling. Degher (2002) reported the take‐back and recycling programmes at Hewlett‐Packard Ltd and concluded that electronic manufacturers and government agencies should work together to better provide customers with environmentally responsible take‐back and recycling programmes. Tsoulfas et al. (2002) performed an environmental analysis of the used SLI batteries sector, based on the logistics involved in the recovery process, and measured the environmental impact of such a process using LCA. Realff et al. (2004) developed a robust‐mixed‐integer linear programming model to support decision‐making for reverse production infrastructure design. Sunthonpagasit and Duffey (2004) examined the engineering economics of crumb rubber facilities and reported that the profitability of a crumb facility appeared to be particularly sensitive to crumb rubber prices, operating costs, and raw material availability.

Nagurney and Toyasaki (2005) developed an integrated framework for the modelling of electronic waste RSCM. They formulated a multi‐tiered, e‐cycling network model (with the objective of profit maximisation) consisting of sources of electronic waste, recyclers, processors, and consumers associated with the demand markets for the distinct products. They provided qualitative properties of the equilibrium electronic waste material flow and price pattern. Listes and Dekker (2005) presented a stochastic programming based approach to a case study on recycling sand from demolition waste in the Netherlands, by which a deterministic location model for product recovery network design was extended to explicitly account for the uncertainties. Wright et al. (2005) illustrated how to achieve an improvement in recyclability of fibre optic cable in a practical and relatively easy way, leading to both environmental and economic benefits. González‐Torre and Adenso‐Díaz (2006) studied the relationships with suppliers and customers from the perspective of environmental demands on the part of the packaging and bottling companies that used glass material.

A general model, based on the principles of reverse logistics, was developed (Logožar et al. 2006) and applied with the aim of reducing the extent of internal aluminium scrap transportation required between certain production units of an aluminium manufacturing plant. Pati et al. (2006) presented a linear optimisation model for the paper industry to compare total system cost of wood, as a raw material, with recycling of waste paper. Pagell et al. (2007) presented a framework that highlighted the supply chain implications for firms forced into EoL product management where recycling was the only viable option. In Biehl et al. (2007), the carpet reverse supply chain was modelled using a simulation package. Pati et al. (2008) formulated a mixed integer goal programming model to assist in proper management of the paper recycling logistics system in India and studied the inter‐relationship between multiple objectives (with changing priorities) of a recycled paper distribution network.

4.2.3 Remanufacturing

Products are dismantled and their parts are used in the manufacturing of the same products (remanufacturing) or in different products (retrieval). The aim of remanufacturing is to bring the product into ‘as new’ conditions by carrying out the necessary disassembly, overhaul, and replacement operations. Industries or products that typically apply remanufacturing include automobiles, electronics and tyres.

Hoshino et al. (1995) defined remanufacturing as recycling‐integrated manufacturing. Thierry et al. (1995) reported that remanufactured products had the same quality as a new product and were sold with the same warranty. They also reported that a copier manufacturer handled the recovery operations in‐house whereas all parts fabrication activities were outsourced. Van der Laan and Salomon (1997) proposed a hybrid manufacturing/remanufacturing system with stocking points for serviceable and remanufacturable products. Ferrer (1997a) examined the case of re‐treading in, the tyre remanufacturing process, and the recommendation of a simple decision rule for selecting the number of times a tyre should be retreaded to maximise its utilisation. Guide (2000) identified and described seven complicating characteristics of production planning and control activities for remanufacturing firms in the United States. The impact of remanufacturing in the economy was studied by Ferrer and Ayres (2000), and, more fundamentally, by Sundin and Bras (2005) who provided arguments for why used products should be remanufactured; revealing that cleaning and repairing were the most critical steps in the remanufacturing process. In a practical case study for remanufacturing, Wendy and Chris (2001) attempted to quantify the life cycle environmental benefits achieved by incorporating remanufacturing into a product system, based on a study of Xerox photocopiers in Australia.

Teunter and Vlachos (2002) studied a single item hybrid production system with manufacturing and remanufacturing. They investigated a variety of cases with different demand, return, manufacturing, and remanufacturing characteristics using simulation. Souza and Ketzenberg (2002) investigated a make‐to‐order remanufacturing system using a queuing network approach. Their focus was to determine the manufacturing–remanufacturing mixture in order to maximise the profit from a strategic perspective. Ferrer and Guide (2002) confirmed that a pure ethical reason could never be the main motive for remanufacturing. Furthermore, they argued that companies need to focus on investing in their core‐activities, rather than offering eco friendly products as part of ethical responsibility. Guide et al. (2003) considered the case of a cellular phone remanufacturing company that acquired used phones with different quality levels, remanufactured them to a single quality level, and sold them at a certain price. The objective was to determine the optimal acquisition prices for used phones and the selling price for remanufactured phones in order to maximise the profit function.

Aras et al. (2004) formulated a model to analyse the conditions under which quality‐based categorisation led to cost savings in hybrid remanufacturing and manufacturing systems. Guide et al. (2005) analysed the performance of static priority rules for a remanufacturing shop that handled two remanufacturable products. Mont et al. (2006) identified a new product group of interest for developing a product‐service system for baby prams. Kim et al. (2006a) discussed the remanufacturing process of reusable parts in reverse logistics, where the manufacturer had two alternatives for supplying parts: either ordering the required parts from external suppliers; or overhauling returned products to bring them back to ‘as new’ conditions. Lebreton and Tuma (2007) presented a real life case study (tyre remanufacturing/re‐treading) with the objective of maximising the profitability of remanufacturing operations. Webster and Mitra (2007) studied a two‐period model of an OEM and an independent remanufacturer where the OEM did not engage in remanufacturing. Mitra (2007) developed a pricing model in the context of remanufactured cellular phones in India to maximise the expected revenue from the recovered products by assuming that probability of selling was a linear function of price.

4.2.4 Repair

The purpose of repair is to return failed products to working condition (possibly with some quality loss) and the purpose of refurbishing is to bring used products up to a specified quality. Most remanufacturing literature also dealt with repair/refurbish (Craig Smith et al. 1996, Ayres et al. 1997, Ferrer 1997a, b, 2001).

In Mabini and Gelders (1991), a multi‐item repairing system was proposed to optimise item repair problems with constraints in terms of in‐stock service levels. Moffat (1992) provided a brief summary of a Markov chain model for analysing the performance of repair and maintenance policies of aircraft engines in the Royal Air Force. Dawe (1995) indicated that customer returns for warranty repair were not being processed and sent back to customers within the agreed upon time. Amini et al. (2005) discussed the competitive value of service management activities, particularly repair services, as well as the importance of the supporting role of effective reverse logistics operations for the successful and profitable execution of repair service activities. A binary integer programming model was developed, in which each binary decision variable indicated whether a given service crew domicile was capable of providing repair service coverage. They presented a case study of a major international medical diagnostics manufacturer.

4.3 Common issues in product recovery

In order to perform product recovery activities such as remanufacturing, recycling, repair and reuse, it is essential to address issues such as collection of returned items, inspection or separation of reusable products and disassembly scheduling.

4.3.1 Collection/acquisition

Collection is the first stage in the product recovery process. Managing the collection and acquisition of used and/or returned products potentially accounts for a significant part of the total costs of any reverse and closed‐loop supply chain. Collection refers to all activities rendering used products available and physically moving them to some point where further treatment is conducted. Collection of used carpet from carpet dealerships (Ammons et al. 1999), take‐back of used copiers from customers (Krikke et al. 1999b), waste paper collection (Pati et al. 2008), used tyres (Kannan et al. 2008b) are typical examples. In general, collection may include purchasing, transportation and storage activities.

Bohm (1997) argued that deposit–refund can be used by regulators as an effective policy tool in a wide range of industries. Guide and Jayaraman (2000) reported that adopting a proactive approach to the implementation of a used product acquisition strategy, by offering the correct amount of incentive to product holders, was of great importance for a company to ensure the sufficient recovery of used products for remanufacturing. In terms of incorporating the variability of individuals' choices in the analysis of deposit–refund systems, Kulshreshtha and Sarangi (2001) constituted the most relevant study in the economics literature. The authors, however, were focused on the use of deposit–refund systems as a price discrimination mechanism rather than the optimal design of collection facility networks. Guide and Van Wassenhove (2001) called on the industry to adopt a proactive approach to used product acquisition, rather than passively accepting the returns. Deposit–refund constitutes an effective means for the firm (as well as the government) to influence the quantity, timing, and possibly the quality of returns.

A deposit–refund system requires consumers to pay a certain deposit at the time of purchase that is refunded upon the return of the used product. Such systems have been commonly used in promoting return and reuse of product packages and containers, e.g. aluminium cans and glass bottles. Other examples of government‐initiated, deposit–refund systems have included car batteries and tyres (Raymond, 2001). Savaskan and Van Wassenhove (2006) analysed the trade‐offs between centralisation and decentralisation of the product collection activity. Wojanowski et al. (2007) presented a continuous modelling framework for designing a drop‐off facility network and determined the sales price that maximised the firm's profit under a given deposit–refund. Kara et al. (2007) presented a simulation model of reverse logistics networks for collecting EoL appliances in the Sydney metropolitan area and calculated the collection cost in a predictable manner. Krikke et al. (2008) presented an approach to optimise the collection (transportation) of dismantled materials from EoL vehicles in real life cases for automobile recycling in the Netherlands.

4.3.2 Inspection/separation

Inspection and separation denotes those operations that determine whether a given product is in fact reusable and to what extent. Thus, inspection and separation result in splitting the flow of used products according to distinct reuse (and disposal) options. Examples are distinguishing repairable and recyclable subassemblies of copiers (Krikke et al. 1999b), inspection of sieved sand for pollution (Barros et al. 1998), and separation of non‐relevant waste paper (Pati et al. 2008). Inspection/separation may encompass disassembly, shredding, testing, sorting, and storage steps.

4.3.3 Disassembly/scheduling

Disassembly is a systematic method of separating a product into its constituent parts, components, subassemblies or other groupings and it is also used to remove toxic elements. It may involve dismantling, demolition or reprocessing. Extensive research work on disassembly scheduling and production planning has been carried out by Gupta and his research group (Gupta and Taleb 1994, 1996, Gupta and McLean 1996, Gupta and Isaacs 1997, Isaacs and Gupta 1997, Gungor and Gupta 1997, 1998, 1999, 2001, 2002, Moyer and Gupta 1997, Taleb et al. 1997, Taleb and Gupta 1997, Zeid et al. 1997, Moore et al. 1998, 2001, Korugan and Gupta 1998, Veerakamolmal and Gupta 1998, 1999, 2002, Boon et al. 2002, Pochampally and Gupta 2003, Gungor 2006, Barba‐Gutiérrez et al. 2008). De Ron and Penev (1995) proposed an approach to determine the degree of disassembly at a single point in time. Mok et al. (1997) described the use of information and new technologies to improve processes in the reverse chain for those situations in which products and equipment need to be disassembled. Gungor and Gupta (1997) and Failli and Dini (2001) proposed heuristic approaches for disassembly planning for the optimisation of recycling processes. Veerakamolmal and Gupta (1998) presented a model to fulfill a demand for components by disassembling different models of returned products. Veerakamolmal and Gupta (1999) discussed a technique for analysing the design efficiency of electronic products, in order to study the effect of EoL disassembly and disposal on the environment. Meacham et al. (1999) included fixed costs to enable the disassembly of a certain item, and centred their approach on determining a cost‐justifying volume. Viswanathan and Allada (1999) and Tang et al. (2001) proposed that, by using group technology; disassembly efficiency could be improved by taking into account the similarities in the operations to be carried out on each product.

Kuo et al. (2000) presented a graph‐based, heuristic approach to perform disassembly analysis for electromechanical products. Pan and Zeid (2001) considered several examples of disassembling products such as a lamp, a car, a window fan, and a two‐stroke engine. Veerakamolmal and Gupta (2002) applied learning algorithms to the disassembly of electronic devices consisting of different configurations of the same modules. Torres et al. (2004) described the process of obtaining a non‐destructive automatic disassembly system for personal computers. Inderfurth and Langella (2006) addressed the disassemble‐to‐order problem, where the yields of disassembly were stochastic, by using heuristic techniques. Langella (2007) used an integer programming model to minimise the total relevant cost incurred over the planning horizon.

Scheduling is the allocation of resources over time to perform a collection of tasks with an objective to find the optimal sequence of jobs. Penev and de Ron (1996) presented models to decide the disassembly sequence and routing, with the aim of minimising the operational costs while fulfilling the production due date. Lambert (1997) presented a graph‐based method for determining the optimum sequence for selective disassembly of discarded complex products. Guide et al. (1997) developed scheduling policies for remanufacturing. Taleb et al. (1997) considered the disassembly scheduling problem for complex product structures with parts and materials commonality. Taleb and Gupta (1997) developed a heuristic method for demand‐driven disassembly planning. Chang et al. (1997) suggested that short‐term planning of vehicle routing and scheduling problems would be a valuable subsequent analysis after the completion of long‐term regional planning for solid waste management.

Johnson and Wang (1998) presented a systematic procedure for generating an optimal disassembly sequence based on maximising the profits of material recovery taking into account material compatibility, clustering for disposal and concurrent disassembly operations. Lambert (1999) proposed a method for solving general sequence generation problems (including the combined disassembly/ clustering problem) by means of linear programming. Gungor and Gupta (2001) proposed a branch and bound algorithm for obtaining approximate optimal disassembly sequences. Lambert (2002) proposed a linear programming model which requires prior determination of all feasible subassemblies and transitions. Brander and Forsberg (2005) developed a lot‐scheduling heuristic for disassembly processes with sequence‐dependent set‐ups.

Lambert (2006) presented a simplified integer linear program and an iterative solution procedure for the case of sequence dependent costs and simple precedence relations. González and Adenso‐Díaz (2006) presented a scatter search meta‐heuristic which deals with the optimum disassembly sequence problem for the case of complex products with sequence‐dependent disassembly costs. Al‐Anzi et al. (2007) studied two deterministic scheduling problems in the computer service provider industry by combining batching and linear time deterioration features. Andrés et al. (2007) proposed a two‐phase approach to determine the optimal disassembly sequence when the disassembly system had a cellular configuration.

MRP for product recovery

Material Requirement Planning (MRP) is a well established and widely used production planning procedure. It is a concept for scheduling production requirements in order to match the demand. Gupta and Taleb (1994) proposed an algorithm (RMRP algorithm) for scheduling the disassembly of a discrete and well‐defined product. Panisset (1998) pointed out that ‘most MRP logic (and the supporting bill of materials) did not provide facilities to plan disassembly’. The problem of best facility location and optimal capacity of the facilities to achieve optimal response time in the lead time study was presented by Bogataj and Bogataj (2004). Kim et al. (2006b) suggested a two‐phase heuristic for disassembly scheduling for multiple product types, with parts commonality, to minimise the sum of set‐up, disassembly operation and inventory‐holding costs using an algorithm incorporating linear and dynamic programming. Barba‐Gutiérrez et al. (2008) presented a methodology to include lot‐sizing in reverse material requirement planning for scheduling disassembly of products.

4.4 Inventory management

A second key area in reverse logistics is inventory management. The producer meets demand for new products and receives used products returned from the market. He has two alternatives for fulfilling the demand. Either he orders the required raw materials externally and fabricates new products or he overhauls old products and brings those back to ‘as new’ conditions. The objective of inventory management is to control external component orders and the internal component recovery process to guarantee a required service level and to minimise fixed and variable costs. The proposed inventory models mainly differ with respect to assumptions on demand and return processes and on the recovery process. A major classification can be made into deterministic versus stochastic models. Diaz and Fu (1997) studied a two‐echelon, repairable item inventory model with limited repair capacity.

4.4.1 Deterministic models

In deterministic models, demand and returns are known in advance for every point in time (Richter and Sombrutzki 2000, Richter and Weber 2001). The objective is to strike an optimal trade‐off between fixed setup costs and variable inventory holding costs. This corresponds to the mindset of the basic EOQ formula in classical inventory theory. Several authors have proposed modifications to this formula taking return flows into account.

Single product deterministic EOQ‐type reverse logistic models were analysed by Schrady (1967), Nahmias and Rivera (1979), Richter (1996a, b), Richter and Dobos (1999), and Koh et al. (2002). Schrady (1967) proposed a deterministic model to manage inventories of repairable items under the condition of fixed lead‐times for external orders and recovery. The author calculated optimal lot sizes in his model. Richter (1996a, b) proposed a model with a different inventory control policy, where the optimal control parameter values were searched; he also discussed their dependence on return rates. Richter (1997) examined the optimal inventory holding policy when the waste disposal (return) rate was a decision variable. A multi product extension of these models was investigated by Mabini et al. (1992). Toktay et al. (2000) addressed the procurement of new components for recyclable products in the context of Kodak's single‐use camera. They modelled the system as a closed queuing network and a heuristic procedure was developed for estimation and control. Minner (2001) combined the problem of safety stock planning in a general supply chain with the integration of external and internal product return and reuse.

Majumder and Groenevelt (2001) studied a two‐period model with one OEM and one independent remanufacturer, and investigated the impacts of alternative allocation mechanisms for returns. Teunter and Van der Laan (2002) presented a specific reverse logistics discounted cash flow (DCF) model and its average cost (AC) approximation with modified holding cost rates. The objective was to minimise the total discounted cost in period (0, ∞). In the paper of Dobos (2003), a generalised Holt–Modigliani–Muth–Simon model was analysed with a quadratic cost structure and a two‐store (first store satisfied the demand, where the manufactured and remanufactured items were stored and the second store collected the returned products which were either remanufactured or disposed) reverse logistics model with continuous disposal was examined. Vlachos and Dekker (2003) analysed the effect of return flows in estimating the initial order quantity of single‐period, random demand products.

Kiesmüller and Minner (2003) extended the determination of optimal order quantities to the multi‐period case with significant recovery and production lead times. Ferrer (2003) dealt with a single‐period problem which examined a remanufacturing system with one collection site, where the yield of returned products was a random variable but demand was known. Ferrer and Ketzenberg (2004) extended the model of Ferrer (2003) to the multiple parts per product – infinite horizon case. The main objective was to study the impact of the relationship between the timing of yield realisation and the procurement lead‐time of new parts on the efficiency of the reverse supply chain. Dobos and Richter (2004) investigated a production‐recycling system where two types of models were analysed. The first model examined the EOQ related costs and minimised the relevant costs. The second model generalised the first model with the introduction of the cost function with linear waste disposal, recycling, production and buy back costs. Robotis et al. (2005) examined procurement and production decisions in a single‐period remanufacturing setting with two collection sites. Mukhopadhyay and Setoputro (2005) developed a profit maximisation model to jointly obtain optimal policies for return policy and modularity level in terms of certain market reaction parameters.

Debo et al. (2005) studied a multi‐period model with one OEM and one or more independent remanufacturers. Ferrer and Swaminathan (2006) studied a firm that makes new products in the first period and competes with a remanufacturer in the second period by selling both new and remanufactured product. The authors analysed two and multi‐period scenarios to assess the effect of first period price on second period profitability. Mostard and Teunter (2006) analysed a newsboy problem with resalable returns. They derived a simple closed form equation that determined the optimal order quantity for a single period inventory (newsboy) problem with returns. Dobos and Richter (2006) investigated a production‐recycling model with quality consideration. By minimising the total EOQ and non‐EOQ related costs, it was shown that it was better to ‘outsource’ the quality control and repurchase only reusable products. Tang et al. (2007) developed models for analysing a disassembly‐remanufacturing system in a real world engine remanufacturing operation, where production was driven by customer orders, and also investigated how the disassembly yield influenced the system performance. Çorbacıoğlu and Van der Laan (2007) analysed a two‐product system with joint manufacturing and remanufacturing in a deterministic setting. In such a system the end product stock contained both manufactured and remanufactured products of different quality.

4.4.2 Stochastic models

Here inventory models which treat demands and returns as stochastic processes are considered. The stochastic inventory models are considered for repair systems and product recovery systems. Literature that considered reverse product flows belonged to the fields of warranty and service parts logistics. The primary consideration was to strike a good balance between inventory holding costs and good customer service and, therefore, most of this literature focused on stochastic inventory models (Cohen et al. 1997, 1999, Murthy and Djamaludin 2002). Listes and Dekker (2005) considered a stochastic programming approach to a case study on recycling sand from demolition waste previously reported by Barros et al. (1998). Wang et al. (2007) considered a supply chain with one supplier, one B2C firm and multiple distribution centres to jointly study supply chain location and inventory policies when product returns were allowed. A new location‐inventory policy based on B2C electronic markets in China was proposed and modelled as an integer bi‐level programming problem.

Periodic review policy

Attention has been focused mainly on deriving optimal control policies under various assumptions and minimising expected costs over a finite planning horizon. Examples of periodic review models are: a model in which returned products can be reused directly (Ferrer 1997b); a model with variable set‐up numbers (Richter 1996a); and models considering the effects of non‐zero lead times (Inderfurth and Van der Laan 2001).

Simpson (1978) considered the trade‐off between material savings due to reuse versus additional inventory carrying costs and proved optimality of a three parameter critical number policy to control order, repair, and disposal for the discrete time case without fixed costs and lead times. Cohen et al. (1980) investigated a dynamic inventory system where both recoverable and serviceable inventories were considered. Cho and Parlar (1991) proposed a multi‐unit inventory control system in which recoverable inventory was allowed to coincide with serviceable inventory, considering that returned products were reused directly. In reality, their model could be regarded as a simple stochastic inventory model with a simplifying assumption that the product issued in a given period was returned with a constant returned rate after a fixed lead‐time. Inderfurth (1997) showed this policy to be optimal also in the case of fixed and identical lead times for repair and procurement. Buchanan and Abad (1998) assumed that returns were a stochastic fraction of the number of items in the market for each period, which was equivalent with an exponentially distributed market sojourn time. The authors derived an optimal procurement policy depending on two state variables, namely the on hand inventory and the number of items in the market.

Van der Laan et al. (1999a, b) presented a detailed analysis of different policies to control serviceable and recoverable stock. They took into account non‐zero lead times for both sources. In particular, a push‐ and a pull‐driven recovery policy were considered. Teunter et al. (2000) compared five methods for setting the holding cost rates for non‐serviceable, remanufactured, and manufactured items, in an average cost (AC) inventory model with reverse logistics. Vlachos and Tagaras (2001) analysed a periodic review inventory system with regular and an emergency supply mode; where policies of the base‐stock type were used at both supply channels, by taking the capacity of the emergency channel into account. They examined two alternative ordering policies: ‘early‐ordering policy’ and ‘late‐ordering policy’. Kiesmüller and Van der Laan (2001) developed a periodic review inventory system for a single reusable product, in which the random returns depend explicitly on the demand stream. Inderfurth et al. (2001) addressed the stochastic remanufacturing problem with multiple reuse options. They derived a periodic‐review optimal policy for allocation of reusable items to different remanufacturing and disposal options under stochastic demands and returns.

Kiesmüller and Scherer (2003) provided a method for the exact computation of the parameters which determined the optimal periodic policy. They provided two different approximations especially in the case of dynamic demands and returns. One was based on an approximation of the value function in the dynamic programming problem while the other approximation was based on a deterministic model and discussed the performance of the approximations by means of numerical examples. Mahadevan et al. (2003) focused on production control and inventory management in the remanufacturing context. They employed a ‘Push’ policy that combined two decisions: when to release returned products to the remanufacturing line; and, how many new products to manufacture. Hahn et al. (2004) dealt with a retailer's operating policies for a perishable product and developed mathematical formulations based on a periodic‐review inventory model under LIFO and FIFO issuing policies. The demand rate was assumed to be a function of the retail price and the proposed model was solved using a Tabu search algorithm.

Continuous review policy

In these models the time axis is modelled continuously and the objective is to find optimal static control policies minimising the long‐run average costs per unit of time. Heyman (1977) considered a continuous review model with the trade‐off relationship between additional inventory holding costs and production cost savings, where product demands and return inter‐occurrence time were assumed to follow respective stochastic processes. Muckstadt and Isaac (1981) developed a model for a one‐warehouse, N‐retailer distribution system with returns where the retailers did not have set‐up costs and they followed a continuous review policy. Van der Laan et al. (1996b) analysed an (s,Q) inventory model in which used products could be remanufactured to new ones. They developed two approximations for the average costs and compared their performance with that of an approximation suggested by Muckstadt and Isaac (1981). Van der laan and Salomon (1997) considered a stochastic inventory system with production, remanufacturing, and disposal operations. They extended the PUSH and PULL strategies to control a system in which all returned products were remanufactured and no planned disposals occur. Korugan and Gupta (1998) for a similar distribution system, made the same assumptions about the retailers and the demand process at the warehouse. They even did not consider the set‐up cost for the serviceable inventory, and developed a model based on an open queuing network with finite buffers.

Minner and Kleber (2001) used Pontryagin's Maximum Principle for finding optimal production and remanufacturing policies for deterministic but dynamic demands and returns when back‐orders were not allowed. Fleischmann et al. (2002) considered inventory control from the perspective of industrial reuse opportunities that have attained growing importance in the rise of environmental concern. They presented a model extending a traditional single‐item poisson‐demand inventory model with a poisson return‐flow of items. Kleber et al. (2002) presented a continuous time, dynamic framework for the product recovery problem with a single return and multiple demand streams for different product variants or qualities. They determined the optimal production, remanufacturing, and disposal policy for a linear cost model by applying Pontryagin's Maximum Principle.

4.5 Waste management

Owing to the growing volume of waste produced, the scarcity of disposal areas and environmental protection, waste management has received increasing attention. Waste management includes incineration and land filling. Waste management is a complex problem that involves taking decisions at strategic and operating levels. Peirce and Davidson (1982) utilised a linear programming model to formulate the optimisation problem of transportation routing among transfer stations, disposal facilities, and long‐term storage impoundments. Jennings and Scholar (1984) formulated the regional hazardous waste management system (RHWMS) as simply a vehicle routing problem in an attempt to accomplish the goal of either minimum cost or minimum risk. Zografos and Samara (1990) dealt only with the problem of a single type of waste to achieve the objectives of minimising transportation risk, travel time, and disposal risk. Koo et al. (1991) used hybrid techniques, including fuzzy theories and multi objective programming models to search for hazardous waste treatment centres in South Korea. Caruso et al. (1993) proposed a multiple objective mixed integer program and a heuristic solution procedure for solving the location‐allocation of waste service users, processing plants, and sanitary landfills with capacity constraints. Chang et al. (1996a, b) and Chang and Wang (1994, 1996) combined the effects of environmental impacts (such as air pollution, noise control and traffic congestion) as a set of risk constraints in an economic‐oriented location model for the solid waste management systems. By characterising the properties of nuclear wastes, Hawickhorst (1997) claimed that the effective management of these nuclear wastes was the prerequisite for the operation of a nuclear energy system, where the operational case of Germany was illustrated. Giannikos (1998) used a multi‐objective model for locating disposal or treatment facilities and transporting waste along the links of a transportation network.

Haastrup et al. (1998) presented a decision support system for urban waste management in a regional area, for evaluating general policies for collection and for identifying areas suitable for locating waste treatment and disposal plants. Hu et al. (2002) presented a cost‐minimisation model for a multi‐time‐step, multi‐type hazardous wastes reverse logistics system. A discrete‐time, linear analytical model was formulated that minimised the total reverse logistics operating costs. Hicks et al. (2004) presented a generic functional model for modelling the material and flow of waste from both a physical and cumulative cost perspective. A study of the physical composition of municipal solid waste, collected in selected Indian cities, by Kumar et al. (2004) revealed that the proportion of plastic and paper in waste generated were very significant and needed immediate attention in order to reduce environmental pollutants. Sheu (2007) presented a coordinated reverse logistics management system of multi‐source hazardous wastes in a given region. A linear multi objective analytical model was formulated that systematically minimised both the total reverse logistics costs and corresponding risks.

Disposal is required for products that cannot be reused for technical or economical reasons. Disposal may include transportation, land filling, and incineration steps. In ReVelle et al. (1991), a synthesised linear programming method was proposed to manage the reverse logistics flows of spent nuclear fuel. Melachrinoudis et al. (1995) developed a multiple objective integer program for the dynamic location of capacitated sanitary land‐fills. Van der Laan et al. (1996a) considered a single‐product, single‐echelon production and inventory system with product returns, product remanufacturing, and product disposal. Bloemhof‐Ruwaard et al. (1996a) presented a two‐level distribution and waste disposal problem, in which demand for products was met by plants; while the waste generated by production was correctly disposed of at waste disposal units. Economics literature provided evidence that deposit–refund is the most preferable policy in terms of the total cost of accomplishing a certain disposal reduction (Palmer et al. 1997, Palmer and Walls 1997). Krikke et al. (1998) presented a comprehensive model to determine an optimal product recovery and disposal strategy for one product type. Ritchie et al. (2000) described a research project carried out within the Manchester Royal Infirmary (MRI) to evaluate and improve the recycling and disposal of pharmaceutical products. Singer et al. (2003) have analysed a game theoretic model in a supply chain with quality and disposable items. A mixed integer linear programming model was developed by Sharma et al. (2007) to facilitate better leasing and logistics decisions (including EoL disposal options) from the perspective of an electronic equipment leasing company.

5. Conclusions and research opportunities

Consumers have become increasingly conscious of their environment and the potential problems that can be created by neglecting it. They have started to show more interest in buying products that are environmentally responsible which includes those that will be taken back by their manufacturers at the end of their useful lives for reuse/repair, remanufacturing and recycling, etc. Environmental legislation and consumer expectations are encouraging manufacturers to design and market environmentally friendly products. Possible cost reductions, more rigid environmental legislations and increasing consumer environmental concerns have led to this increased focus on reverse supply chain management. In view of this focus, this literature review was conducted to analyse the literature on environmental regulation, product recovery, inventory management and waste management related issues. The review categorised the papers according to their content and the classification presented in Section 3. On the basis of this review, the following conclusions and recommendations are presented:

	the majority of articles addressed disassembly/scheduling issues in the contexts of remanufacturing and recycling;
	the study of more complex objective functions that include, for example, inventory holding and shortage costs, presents an interesting opportunity for future research;
	there is a need to develop more robust stochastic models for complex product recovery problems;
	revenue management for recovered products is an important subject that has not yet received sufficient attention;
	future research opportunities include: determining starting points for initial prices offered for used products; generic methods for determining and rating the quality of used products and determining the resulting distributions of nominal quality; and specifying the cost to remanufacture as a function of the price offered (Jayaraman 2006);
	besides the effect on regular orders, the timing and size of emergency orders requires analysis;
	governments can improve the collection rate for used products by taking on the responsibility of operating deposit–refund systems (i.e. the deposit will be kept by the government until the product is returned);
	there is a research potential in the design and testing of lean remanufacturing principles, or the ‘remanufacture‐to‐order’ approach, within OEM remanufacturing (Seitz 2007);
	more countries should implement legal requirements, using the European model, to encourage product return flows; and
	there is an opportunity for research that compares the performance of different inventory policies when considering multiple quality states of the returned items.

Acknowledgements

The authors would like to acknowledge the assistance of the anonymous reviewers and the journal editors who provided suggestions for improvements to the original paper.