This article was downloaded by: [Australian National University]

On: 07 January 2015, At: 19:18

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

International Journal of Sustainable Engineering

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/tsue20

Reverse logistics network design: a conceptual

framework for decision making
a a
Theresa J. Barker & Zelda B. Zabinsky
a
Industrial Engineering , University of Seattle , Washington, USA
Published online: 11 Dec 2008.

To cite this article: Theresa J. Barker & Zelda B. Zabinsky (2008) Reverse logistics network design: a conceptual framework
for decision making, International Journal of Sustainable Engineering, 1:4, 250-260, DOI: 10.1080/19397030802591196

To link to this article: http://dx.doi.org/10.1080/19397030802591196

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 International Journal of Sustainable Engineering
Vol. 1, No. 4, December 2008, 250–260

Reverse logistics network design: a conceptual framework for decision making

Theresa J. Barker* and Zelda B. Zabinsky
Industrial Engineering, University of Seattle, Washington, USA
(Received 14 July 2008; accepted 31 October 2008)

Reverse logistics play a significantly increased role in supply chains. Producers are incorporating reverse logistics
into their supply chain design because of governmental legislation, economic benefits from product recovery, and
customers’ demands for better environmental practices. Motivated producers face two main challenges in reverse
logistics network design: (1) how do you build product recovery activities into traditional forward logistics
networks? and (2) how do you manage the impact of uncertainty in the reverse logistics supply chain? Moreover,
producers need to consider many tradeoffs leading to efficient network design. Published case studies have
discussed these tradeoffs, and we describe a framework to evaluate these tradeoff considerations. We then apply
Downloaded by [Australian National University] at 19:18 07 January 2015

the framework to three new case studies: medical device refurbishing, municipal e-waste recycling, and carpet
fibre recycling.
Keywords: reverse logistics; product recovery; network design; decision making models

1. Introduction With the popularity of product recovery and

recycling functions comes another compelling ques-
Over the past two decades, traditional supply chains
tion: how predictable is the supply chain for reverse
in which product is shipped from producer to
logistics? Because of much higher levels of uncer-
consumer are giving way to contemporary supply
tainty in volume and condition of returned product,
chains which incorporate reverse logistics, or the producers face much higher uncertainties in supply
process of recovering value from end-of-life pro- and demand of returned product, which presents a
ducts. Because traditional supply chains are crucial challenge. Producers incorporating reverse
designed to provide a certain quantity of product logistics into their supply chains need to consider the
to the customer at a certain time, logistics systems impacts of higher uncertainties, and to mitigate those
are typically not designed to accommodate reverse impacts as much as possible.
logistics efficiently. Yet the disposal of waste Most of the reverse logistics work has been
presents a critical challenge. Kirstie McIntyre of product- or industry-specific (e.g. electronics,
Hewlett Packard says that we have a big problem: recycled paper, auto parts, carpet, and copiers),
‘The primary output of today’s production pro- primarily because of efficiencies that may be gained
cesses is waste. Across all industries, less than 10 from product-specific knowledge in recycling or
percent of everything that is extracted from the remanufacturing processes (Realff et al. 2000). The
earth (by weight) becomes usable products. The vast majority of industry-specific models have been
remaining 90 percent becomes waste from produc- quantitative mixed-integer linear programming facil-
tion…’ (McIntyre 2007). ity location-allocation models (Barros et al. 1998,
Waste reduction is at the heart of a new trend, Krikke et al. 1999a, Louwers et al. 1999, Fleischmann
which has three main motivators: governmental et al. 2001). More recently, stochastic programming
legislation, economic value to be recovered in returned or robust optimisation approaches have been added
product, and environmental concerns. The govern- to model inherent uncertainties (Realff et al. 2004,
ment-mandated motivation is especially strong in Listeş and Dekker 2005, Hong et al. 2006).
Europe, where governmental regulations are compel- A number of researchers have used case study
ling businesses to address recovery and disposal of analysis and review of the reverse logistics literature
end-of-life products. In the US, the motivation has to develop reverse logistics frameworks and models.
been largely due to economic factors – recovering Many models have suggested a number of tradeoffs
value where there is an economic incentive – but that must be considered for efficient reverse logistics
legislation has begun to compel producers to set up networks. These tradeoffs form an implicit frame-
systems for product recovery and safe disposal in a work to determine network design decisions, such
number of industries, particularly for electronic as exploiting economies of scale using an industry-
product disposal (also known as e-waste). wide recovery system versus protecting proprietary

*Corresponding author. Email: barkertj@u.washington.edu

ISSN 1939-7038 print/ISSN 1939-7046 online
# 2008 Taylor & Francis
DOI: 10.1080/19397030802591196
http://www.informaworld.com
 International Journal of Sustainable Engineering 251

intellectual information through a company-specific should occur early in the collection process or at a
recovery system. centralised location, and whether to use traditional
In this paper, we describe a framework for supply chain actors (e.g. manufacturers, middlemen
evaluating tradeoffs in network design decisions and retailers) or specialised parties (e.g. secondary
based on key considerations. We analysed 13 case material dealers, recovery facilities, and logistics
studies and abstracted key considerations that were providers).
common among all the case studies. These considera- In 2000, Fleischmann et al. (2000) identified
tions led to one of eight possible configurations. We characteristics of product recovery networks by
then classified an additional 24 case studies in the dividing them into three types (bulk recycling,
literature (De Brito et al. 2003), for a total of 37 case assembly product remanufacturing, and re-usable
studies, to see if the same considerations were valid. items) and then classifying network characteristics
Finally, we developed three new case studies, which within each type (e.g. dedicated facilities, reuse in
represented three configurations within our frame- original market, mandatory recovery). This model
work. A producer considering reverse logistics can was predicated on the type of product recovery
look at the considerations that we identified, along process, such as bulk recycling. Tradeoff choices were
Downloaded by [Australian National University] at 19:18 07 January 2015

with classified case studies, to lead them to a discussed such as centralisation versus decentralisa-
particular configuration of network design decisions. tion of recovery activities, single-activity facilities
The remainder of the paper is organised as versus multiple-activity facilities, and integrated net-
follows. Section 2 contains a literature review. In work routing versus separate network routing.
Section 3, we describe the framework and illustrate it The framework developed by De Brito and
with a selection of published case studies. Section 4 Dekker (2003a) looked at characteristics of reverse
presents our new case studies and their associated logistics systems according to drivers (economics,
configurations in the framework. Discussion and legislation and ‘green citizenship’), return reasons
future research are presented in Section 5. (manufacturing returns, distribution returns and
customers returns), recovery processes (collection,
inspection/selection/sorting and recovery), types of
2. Background and literature review returned products (composition, deterioration and
An early framework was proposed by Flapper (1996). use-pattern), and the actors (forward supply chain,
The framework was an overview of the logistics of specialised reverse supply chain and opportunistic
reuse, in which the author categorised reuse activities players). They stated that the combination of these
into collection, processing, and distribution, and characteristics determines the issues that arise in a
discussed various aspects of each activity. Flapper specific reverse logistics system.
described a number of tradeoff considerations, These frameworks have presented a number of
including whether to collect directly from customers tradeoff considerations. Other frameworks have
or at depots, whether the network should be taken a business analysis approach. In 1998, Carter
geographically wide-spread or localised, whether and Ellram (1998) proposed a framework for reverse
different items should be collected together or logistics from a business perspective. By analysing a
separately, whether to transport return product back wide variety of research papers, they developed a
to a processing plant or to reuse or recycle the framework predicated on drivers (Regulations,
product locally, and whether inspection and sorting Customers, Policy Entrepreneurs, and Uncertainty)
should be done immediately on collection or at the and constraints (Stakeholder Commitment, Top
point of processing. Management Support, Incentive Systems, Quality of
Fleischmann et al. (1997) presented a framework Inputs, and Vertical Coordination). The framework
that is widely cited throughout the literature. Using also distinguishes among internal and external
case studies from practice, the authors identified drivers. For instance, the framework suggests that
characteristics of reverse logistics networks, and the principal internal driver is having at least one
compared them to traditional logistics networks. policy entrepreneur, and that top management
They proposed a framework based on the flow of support, stakeholder commitment, and appropriate
reverse distribution from producer to user and back incentive systems are necessary for successful imple-
to producer, defined types of motivations (govern- mentation of reverse logistics.
ment legislation, economic value, ‘ecological’ image), A decision-making model for third-party logistics
types of disposal (reuse, repair, recycling, remanu- providers (3PLs) was proposed by Krumwiede and
facturing), and discussed the issue of separate versus Sheu (2002), who interviewed 3PLs and analysed
integrated network flows. They discussed tradeoff related research articles. They identified a three-stage
considerations such as whether testing and sorting flow: retrieval (collection), transportation (including
 252 T. J. Barker and Z. B. Zabinsky

storage), and disposition (two types: on-site and off-

site). The authors then developed a model for
decision making that included researching existing
issues and identifying current customers, building
marketing channels using those customers, identify-
ing niches, and performing a feasibility study.
Meade et al. (2007) analysed an exhaustive
number of articles from the literature to determine
the definition, functions and mechanisms of reverse
logistics. The result was a framework composed of
driving forces that are environmental factors (e.g. Figure 1. Flow of reverse logistics activities.
regulation and environmental friendliness), and busi-
ness factors (e.g. liberal customer returns and diverse products and industries. We identified net-
customer satisfaction). They also observed that there work design characteristics and tradeoff considera-
was a need for generalised models in the literature. In tions that were common among these case studies.
Downloaded by [Australian National University] at 19:18 07 January 2015

a similar vein, Rubio et al. (2008) confirmed the need Table 1 lists the eight possible configurations, and
for new research into strategic aspects and organisa- lists each of the 13 case studies under its associated
tional frameworks for reverse logistics, despite configuration.
existing quantitative models and case studies. We further tested the framework using a set of
Often high level decisions, such as where to case studies in De Brito et al. (2003). Of their 67 case
perform recovery activities or whether to use a studies, we examined 31 case studies written in
third-party processor, are not based on a structured English. Of these 31 case studies, seven were
approach to making those decisions. Previous frame- duplicated in our original set of 13, leaving 24
works described tradeoffs in network design, con- additional case studies to be classified. Table 1
stituting an implicit framework for evaluating classifies our complete set of 37 case studies under
tradeoff choices. We present a structured approach, the eight possible configurations.
by providing considerations that lead to one of eight We found that approximately one-third of the
configurations, and by presenting case studies that case studies (12) had the configuration industry-wide
inform the design decisions. A manufacturer con- collection, centralised sort/test and secondary facility
sidering reverse logistics functionality can use our processing. In these systems, returned product is
framework to make those design decisions. collected via an industry-wide system, transported to
a central facility for sorting and testing, and then
processed at a secondary facility which is not
3. Framework and published case studies company-specific. This configuration is common for
In reverse logistics, there are three fundamental a commodity-type recycling system, such as used
stages of flow: collection, sort/test and processing construction sand (Barros et al. 1998) or carpet
(De Brito and Dekker 2003a). Processing may include (Realff et al. 2000).
finished product reuse, refurbishment and spare parts Three case studies (3) had the configuration
recovery, reprocessed raw material and disposal of industry-wide collection, distributed sort/test and
waste. A product recovery flow diagram, adapted secondary facility processing. These case studies were
from Flapper (1996) and De Brito and Dekker recycling studies, two recycled paper systems and an
(2003a) is shown in Figure 1. electronic waste recycling system, in which the return
Each stage of flow has a set of network product was first sorted at the collection site before it
characteristics involving critical tradeoffs, which lead was transported to the recycling facility (Bloemhof-
to choices for network design decisions. The design Ruwaard et al. 1996, Kleineidam et al. 2000, Hong
decisions and their associated tradeoff considerations et al. 2006).
are presented in a framework, shown in Figure 2. The All of the case studies with industry-wide collec-
framework provides a classification system consisting tion systems performed processing at secondary
of eight possible configurations (see Table 1). The key facilities; none used original facilities. In industry-
tradeoff considerations were identified primarily wide collection systems, processing is highly likely to
from 13 published case studies representing a variety occur at a secondary facility, because it is relatively
of industries, as well as previous frameworks in the inefficient to retrieve a specific manufacturer’s
literature. Some case studies came from Fleischmann products from an industry-wide collection system
et al. (2000), a seminal paper in this field, and the rest and then deliver those products to the original
were selected through a literature search to represent manufacturing facility for processing. However, such
 International Journal of Sustainable Engineering 253
Downloaded by [Australian National University] at 19:18 07 January 2015

Figure 2. Framework for network design decisions.

a configuration may become viable if certain condi- Another one-third of the case studies (12) had the
tions change – for instance, if a specific manufac- configuration company-specific collection, centralised
turer’s products were easily identified and separated sort/test and original facility processing. Company-
out from a industry-wide collection stream, then specific systems often have this configuration, in
shipped directly from a collection site to the original which a company retrieves its own products, and
facility to avoid unnecessary transportation costs. processing is done at the original facility. This
One example of such a system would be the use of configuration is consistent with a proprietary refurb-
RFID tagging for a particular manufacturer’s pro- ishing system, such as car engines (Seitz and Peattie
duct in an industry-wide collection system; RFID 2004) or reusable glass soft drink bottles (Del Castillo
tagging would make tracking and retrieval of that 1996).
manufacturer’s products much simpler. Another We found that of the remaining case studies, four
example is the specific labeling of a manufacturer’s (4) had the configuration company-specific collection,
product with a guarantee of end-of-life return by the centralised sort/test and secondary facility processing.
manufacturer, such as for flooring squares, allowing This system makes sense for a proprietary collection
the collector to contact the manufacturer to return system using a third-party logistics provider, such as
the product. toner cartridges (Bartel 1995), for a company without
 254 T. J. Barker and Z. B. Zabinsky

Table 1. Eight configurations in the framework. Table 1. Continued.

1. Industry-wide collection, centralised sort/test, and original facility 2 of 37 case studies

processing Kroon and Vrijens (1995) Reusable packaging
0 of 37 case studies Thierry et al. (1995) Copier refurbishing
No published case studies found
2. Industry-wide collection, centralised sort/test, and secondary space in its original facility, such as business lease-
facility processing return computers (Fleischmann 2000, Fleischmann
12 of 37 case studies et al. 2004) or where the secondary processing is
Barros et al. (1998) Construction sand recycling
dramatically different from original manufacturing,
Chang et al. (2000) Municipal curbside waste
Farrow et al. (2000) Recycled plastic kayaks such as proprietary battery recycling (Yender 1998).
Guide and van Wassenhove (2001) Cellular phone Another four (4) had the configuration company-
remanufacturing specific collection, distributed sort/test and original
Klausner and Hendrickson (2000) Power tool facility processing. Proprietary systems, where sorting
remanufacturing
Krikke et al. (1999b) PC monitor recycling
can be done at the collection site, fall into this category.
Louwers et al. (1999) Carpet recycling For instance, some reusable container systems perform
Downloaded by [Australian National University] at 19:18 07 January 2015

Nagel and Meyer (1999) Refrigerator remanufacturing inspection at distributed depots, before being sent to a
Realff et al. (2000) Carpet recycling new user (Duhaime et al. 2001), or copier manufacturers
Spengler et al. (1997) Steel by-products inspect at disassembly centres before shipping to the
Staikos and Rahimifard (2007) Shoe recycling
Wang et al. (1995) Cardboard recycling
refurbishing facility (Krikke et al. 1999a).
3. Industry-wide collection, distributed sort/test, and original facility Finally, we found two (2) case studies with the
processing configuration company-specific collection, distributed
0 of 37 case studies sort/test and secondary facility processing. In both
No published case studies found
studies, return product was sorted and processed by a
4. Industry-wide collection, distributed sort/test, and secondary
facility processing third party at a decentralised location. One was a
3 of 37 case studies reusable container system (Kroon and Vrijens 1995)
Bloemhof-Ruwaard et al. (1996) Paper recycling and the other was a copier refurbishing system
Hong et al. (2006) e-Scrap recycling (Thierry et al. 1995).
Kleineidam et al. (2000) Paper recycling
5. Company-specific collection, centralised sort/test, and original
If a producer is considering incorporating reverse
facility processing logistics into their supply chain, they can look at the
12 of 37 case studies considerations listed in Figure 2 to identify a specific
De Brito et al. (2003b) Lab equipment restocking configuration. They can then see all the case studies
De Brito et al. (2003b) Mail order company restocking to help them design their network.
De Brito et al. (2003b) Refinery spare parts restocking
Del Castillo and Cochran (1996) Reusable glass soft drink bottles Next we describe the tradeoff considerations in
Dı́az and Fu (1997) Subway spare parts restocking each stage in detail.
Guide and van Wassenhove (1997) Military aircraft
remanufacturing
Linton and Johnston (1999) Circuit board refurbishing 3.1 Stage A: collection
Maslennikova and Foley (2000) Electronic product
Collection systems are either industry-wide, in which
remanufacturing
McGavis (1994) Printer toner cartridge recycling the same type of product from multiple producers is
Meyer (1999) Cosmetics products restocking collected within the system, or company-specific, in
Seitz and Peattie (2004) Car engine remanufacturing which a company collects only its own products for
Toktay et al. (2000) Single-use camera recycling recovery. For company-specific collection systems,
6. Company-specific collection, centralised sort/test, and secondary
producers can use proprietary routing, in which the
facility processing
4 of 37 case studies producer uses its own transportation system for
Bartel (1995) Printer toner cartridge recycling collection, or producers can outsource the collection
Fleischmann (2000) Business computer refurbishing to a third-party logistics provider.
Thomas Jr. (1997) Aircraft engine remanufacturing Industry-wide collection systems tend to be used
Yender (1998) Battery recycling
7. Company-specific collection, distributed sort/test, and original
for commodity-type products, such as paper recycling
facility processing (Bloemhof-Ruwaard et al. 1996). These systems are
4 of 37 case studies becoming prevalent for computers and electronic
Duhaime et al. (2001) Reusable postal containers products, due to government mandates for industry-
Gupta and Chakraborty (1984) Glass scrap recycling
wide, e-waste collection systems (Hong et al. 2006). A
Krikke et al. (1999a) Copier refurbishing
Rudi et al. (2000) Wheelchair refurbishing benefit of this type of system is economies of scale,
8. Company-specific collection, distributed sort/test, and secondary due to higher volumes. It also does not complicate a
facility processing company’s forward supply chain, as an industry-wide
 International Journal of Sustainable Engineering 255

system is typically a completely separate product of higher transportation costs for shipping scrap to
return stream, collected by a third-party entity, as it is the testing facility first, rather than directly to waste
for electronic waste (e-scrap). However, an individual disposal.
company has limited control over this type of Distributed sort/test sites are often used if low-
collection system, and that includes costs and cost testing procedures are available, such as for
routing. Also, higher start-up costs may be incurred paper recycling (Bloemhof-Ruwaard et al. 1996,
for an industry-wide collection system, because of the Kleineidam et al. 2000), machine refurbishing
much larger scale and scope of the system. (Thierry et al. 1995, Krikke et al. 1999a), or reusable
A company-specific collection system has its own containers and equipment (Duhaime et al. 2001,
benefits and drawbacks. It is particularly beneficial Kroon and Vrijens 1995, Rudi et al. 2000). Scrap can
when the company has a strong direct relationship be identified early and shipped to waste disposal,
with its customer, such as a lease-return relationship, reducing transportation costs. However, testing
or when there is high customer trade-in behaviour, procedures must be consistent and reliable, and the
such as there is in the business computer market network may be more complicated because scrap and
(Fleischmann 2000, Fleischmann et al. 2004). The usable return product are shipped in separate
Downloaded by [Australian National University] at 19:18 07 January 2015

company-specific collection system tends to streams.

strengthen those customer relationships, enhancing
marketing and sales efforts. However, transportation
costs may be higher than in an industry-wide 3.3 Stage C: processing
collection system, because a company-specific system Once the type of processing is determined (recycling,
cannot take advantage of economies of scale avail- reprocessing raw material, refurbishment and spare
able to higher volumes that an industry-wide system parts recovery, or reuse), the key decision is whether
would handle. to reprocess at the original facility, which is the
Company-specific proprietary routing is an method for copiers (Krikke et al. 1999a), or at a
attractive choice when a company wishes to protect secondary facility, which is the method for carpet
intellectual and proprietary information. It can also (Realff et al. 2000).
be desirable for integrating forward and reverse Processing at the original facility provides
flows, such as for drop-off and pickup of reusable increased efficiency from use of original facility
containers (Kroon and Vrijens 1995). This system is equipment and processes, and it is often used for
also beneficial when there are relatively few customer machine refurbishing or spare parts recovery proces-
sites. One drawback is potentially higher costs, as sing. However, there may be a need for increased
proprietary routing may be more expensive than processing capacity, which would be a drawback.
outsourcing the collection system. The benefits of processing at a secondary facility
Company-specific third-party routing is better include economies of scale if done across the entire
positioned to take advantages of economies of scale, industry rather than for a single manufacturer, which
as third-party logistics providers can pool shipping makes this a good choice for a bulk commodity-type
and facilities needs for multiple customers. This type product such as construction sand. The drawbacks
of system may also be preferable for companies with include the need to establish new, separate facilities
large numbers of customer sites. Nevertheless, a with a possible loss of processing efficiency.
third-party routing system has the drawback of
reduced control by an individual company when it
comes to intellectual and proprietary information. 4. Three new case studies
We developed three new case studies: (1) a municipal
e-waste recycling program, (2) a medical device
3.2 Stage B: sort/test
refurbishing system, and (3) a carpet fibre recycling
Sorting and testing can be performed either at a program. Each case study represents a different
centralised site, or at distributed locations. A configuration.
centralised site is common for a commodity-type
product, such as construction sand recycling (Barros
et al. 1998) or carpet recycling (Louwers et al. 1999, 4.1 Municipal electronic waste recycling
Realff et al. 2000), owing to efficiencies from higher The City of Bellevue, Washington, initiated a curb-
volumes. But a centralised site is also desirable for side waste disposal system for electronics products.
high-cost testing procedures, because it minimises Bids were solicited for a contract with the munici-
costs of testing equipment and specialised labor. One pality to collect electronic products from residential
drawback to centralised sorting and testing is the risk customer sites and to dispose of them through a
 256 T. J. Barker and Z. B. Zabinsky

commercial electronics recycler. The contract was transportation to the recycling facility, greater
awarded to Allied Waste, which collects the electronic efficiencies in disassembling and recycling, and
waste and delivers it to Total Reclaim, a large greatly mitigate the impacts of variability in
electronics recycler in an industrial section of volume.
Seattle (see Figure 3).
This system’s major challenge is that it is reactive
State legislation was passed in Washington in
to product design, and since the system is an ‘end-of-
2006 mandating that producers of electronic products
provide a system for disposal of those products in an pipe’ solution, it cannot take advantage of efficiency
environmentally safe manner, which will take effect improvements through product design.
in January 2009. However, the City of Bellevue Framework configuration:
decided to implement its program proactively in Collection Sort/test Processing
advance of the legislation, in order to improve service Industry-wide Centralised sort/ Secondary facility
to its customers. collection test processing

Electronic waste is collected at the customer site, Because electronic waste is a commodity-type
and then transported to a holding area within the product that is produced by multiple manufacturers,
Downloaded by [Australian National University] at 19:18 07 January 2015

city. The e-waste is placed into gaylords, bulk cargo industry-wide collection is preferred, and a secondary
containers, which are then loaded into a semi-trailer; facility – an electronics recycler – is used for
when the trailer is full (about 16–18 gaylords), Total processing. While it may be possible for individual
Reclaim recycling picks up the trailer and delivers it
electronics manufacturers to set up their own
to their electronic processing facility.
company-specific collection system to process pro-
The electronic waste is sorted at the facility.
ducts at their original facility, the fact that a majority
Products are disassembled – plastic shredded, heavy
of electronics manufacturing facilities are located
metals and glass retrieved – and all components are
outside the US would make a proprietary system
recycled.
much more costly.
This system has the following advantages:
The City of Bellevue system takes greater
N Proactive approach – ‘ahead of the curve’. The advantage of economies of scale than prior electro-
program was implemented well in advance of nics products recycling systems, such as the e-Scrap
government legislation, allowing the possibility case study (Hong et al. 2006). In the e-Scrap case,
of a more favourable contract negotiation for customers brought used electronic products to a
the city. municipal collection site, where the products were
N Encourages high customer compliance. Pre- sorted and then sent to commercial processing sites.
viously, electronic waste recycling programs Costs can be reduced in the City of Bellevue case
have relied on customers to bring their electronic because the electronic recycling is integrated into the
waste to a central or regional recycling location; solid waste stream.
the convenience of curbside waste pickup
encourages customer compliance.
N Economies of scale. Because the program is 4.2 Medical device refurbishing
part of a municipal waste disposal stream, it A major medical device manufacturer on the West
exploits economies of scale. For example, Coast has a refurbishing program, which operates on
higher volumes mean full truckloads for a trade-in basis with customers under a service

Figure 3. City of Bellevue municipal curbside e-waste recycling.

 International Journal of Sustainable Engineering 257

Figure 4. Major medical device manufacturer refurbishing.

Downloaded by [Australian National University] at 19:18 07 January 2015

contract. An outdated machine may be This system’s major challenge is uncertainty in

either shipped back to the manufacturer’s facility supply and demand. The volume and condition of
for refurbishing, or shipped directly to a national return products is highly variable, complicating sales
electronics recycler in Chicago, Sims/URI (see forecasting and inventory control of used machines.
Figure 4). The producer’s holding warehouse provides some
At the customer site, the outdated product is mitigation as an inventory buffer.
evaluated for recycling or possible refurbishing. Framework configuration:
Products to be recycled are shipped to the electronics Collection Sort/test Processing
recycler directly, reducing transportation costs that Company-specific Distributed sort/ Original facility
would be incurred if the product was shipped first to collection test processing
the manufacturer’s warehouse on the West Coast and The medical device manufacturer’s system involves
then to the recycling facility. Products that may be a single manufacturer’s product in which proprietary
refurbished are shipped to a warehouse near the knowledge is critical; company-specific collection and
manufacturing plant. They are held until a customer original facility processing are preferable. This case has
order is received, and then they are refurbished to fill the advantage of distributed sort/test, shipping scrap
the order, avoiding the investment of unnecessary directly from the customer site to the recycler, saving
cost to refurbish a product that does not get sold. transportation costs. Performing sorting and testing at
Periodically, the manufacturer’s warehouse inventory the collection site, as in this case study, is preferable over
is culled for excess inventory, and redundant centralised sorting and testing whenever possible.
machines are sent to the electronics recycler. Dealing with pre-sorted returned product reduces costs
The electronics recycler has the capability to at the processing centre, as well as transportation costs,
recover spare parts for the manufacturer’s service for product that cannot be processed. For example, in
part inventory. This provides valuable spare parts to the business lease-return computer case study
repair older machines still in use by customers, (Fleischmann 2000, Fleischmann et al. 2004), all return
typically those machines which are under a service computers came into a central facility before they were
contract for seven to 10 years. recycled or refurbished, unlike the system in this new
This system has the following advantages: case study.
N On-site evaluation for sort/test. Transporta-
tion costs are minimised by avoiding shipping 4.3 Carpet fibre recycling
scrap to the manufacturer’s warehouse before Shaw Industries, a Dalton, Georgia flooring manu-
shipping to the recycler. facturer, collects used carpet from a number of
N Refurbish on customer order. Because ma- individual third-party recycling centres located
chines are refurbished to order, unnecessary nation-wide. The carpet is sorted at the recycling
costs are avoided that would be incurred if a centres for a specific type of fibre (nylon 6), then
machine was refurbished and remained unsold. bundled and shipped to Shaw’s secondary nylon 6
N Flexible spare part recovery. Spare parts can recycling plant in Georgia (see Figure 5).
be recovered either by the recycler or at the Used carpet is tested at the third-party recycling
manufacturing plant, allowing flexibility in centre with a hand-held device that identifies the fibre
maintaining outdated machine spare parts. content. Nylon 6 carpet is separated from other carpet
 258 T. J. Barker and Z. B. Zabinsky

Figure 5. Shaw Industries carpet fibre recycling.

types and bundled by the recycler, then shipped to a Shaw’s carpet fibre recycling system is an
Downloaded by [Australian National University] at 19:18 07 January 2015

warehouse near the Shaw recycling plant. The industry-wide collection system with the advantage
carpet is unbundled and tested again; if a bundle of distributed sort/test. Although the recycled fibre is
is 96% or greater nylon 6 fibre, a bonus payment is used in its original manufacturing facility, the
given to the recycler. It is then stored in the recycling processing is performed at a secondary
warehouse both to adjust the humidity level of the facility, which is typical for an industry-wide collec-
carpet, and to buffer seasonal variations in inven- tion system. The distributed sort/test feature is an
tory levels. improvement over the prior carpet case study listed in
Used carpet is processed at the depolymerising the paper by Realff et al. (2000). The advantage of
plant, producing raw caprolactum fibre, the basis for testing at the collection site is that the returned
nylon 6 fibre. The quality of the caprolactum fibre is product is pre-sorted, avoiding excess transportation
actually higher than fibre obtained elsewhere by costs for non-nylon 6 carpet.
Shaw for carpet production, and it can be sold by
Shaw to other carpet manufacturers. Shaw’s custo-
5. Discussion and future research
mers indicate that, given a comparable price point
between products made from recycled fibre and those Reverse logistics complicates the supply chain. Often
made from virgin fibre, they will consistently buy the the reverse flow network is added as a solution to
recycled product. end-of-life product returns, and does not take
This system has the following advantages: advantage of efficiencies in the existing supply chain.
A better understanding of the tradeoffs inherent in
N Collection site testing incentive program. Over network design decisions is essential for producers
the years, Shaw has found that the content of and industries to develop efficient reverse logistics
baled carpet bundles is consistently above 98% networks.
nylon 6, due to their incentive program. In this paper, we identified key tradeoff con-
N Customer preference for recycled product. siderations for network design decisions, adapted
Having a recycling process for carpet produc- from prior frameworks and by case study analysis of
tion translates to increased customer sales. 13 case studies. We classified an additional 24 case
N Economies of scale. Having an industry-wide, studies using our framework to evaluate the tradeoff
nation-wide collection program helps promote considerations, and we applied the framework to
efficiencies in transportation and processing of three new case studies. The identification of network
used carpet. configurations with tradeoff considerations provides
This system’s major challenge is capital invest- a structured approach for new manufacturers to
ment and costs of production. Under current condi- begin the process of designing their reverse logistics
tions, production costs of recycled fibre slightly networks. Our classification of 37 case studies and
exceed the cost to purchase raw material. However, three new case studies provides examples across
changing conditions, such as an increase in the cost of products and industries for new manufacturers to
oil, may make recycled fibre more cost-effective. learn from.
Framework configuration: Future work will include a decision making model
Collection Sort/test Processing that integrates quantitative detailed models with high
Industry-wide Distributed sort/ Secondary facility level configuration decisions to quantity the tradeoffs
collection test processing and assess the efficiency of network design, using
 International Journal of Sustainable Engineering 259

multi-objective programming and probabilistic Fleischmann, M., 2000. Quantitative Models for Reverse
approaches for managing uncertainties. Logistics. Thesis (PhD). Erasmus University Rotterdam.
Fleischmann, M., et al., 1997. Quantitative models for
reverse logistics: a review. European Journal of
Acknowledgements Operational Research, 103, 1–17.
Fleischmann, M., et al., 2000. A characterisation of
The authors wish to thank the referees for their constructive
logistics networks for product recovery. Omega, 28,
comments and suggestions. We also gratefully acknowledge
653–666.
the valuable collaboration and assistance provided by John
Fleischmann, M., et al., 2001. The impact of product
Aanenson and Tom Spille of City of Bellevue, Don Frey of
recovery on logistics network design. Production and
Allied Waste, and John Conyers, Jr., and Rick Ramirez of
Operations Management, 10 (2), 156–173.
Shaw Industries Group.
Fleischmann, M., et al., 2004. Chapter 4: reverse logistics
network design. In: R. Dekker, et al., eds. Reverse
References Logistics Quantitative Models for Closed-Loop Supply
Chains. Berlin: Springer-Verlag.
Barros, A.I., Dekker, R. and Scholten, V., 1998. A two- Guide, Jr., V.D.R. and van Wassenhove, L.N., 1997.
level network for recycling sand: a case study. European Rough-cut capacity planning for remanufacturing firms.
Downloaded by [Australian National University] at 19:18 07 January 2015

Journal of Operational Research, 110, 199–214. Production Planning and Control, 8 (3), 237–244.
Bartel, T., 1995. Recycling program for printer toner Guide, Jr., V.D.R. and van Wassenhove, L.N., 2001.
cartridges and optical photoconductors. In: Proceedings, Managing product returns for remanufacturing.
IEEE International Symposium on Electronics and the Production and Operations Management, 10 (2), 142–155.
Environment, 1–3 May 1995, Orlando, FL. 225–228. Gupta, T. and Chakraborty, S., 1984. Looping in a
Bloemhof-Ruwaard, J.M., et al., 1996. An environmental multistage production system. International Journal of
life cycle optimization model for the European pulp and Production Research, 22 (2), 299–311.
paper industry. Omega, 24, 615–629. Hong, I.-H., et al., 2006. Planning the e-scrap reverse
Carter, C.R. and Ellram, L., 1998. Reverse logistics: a review production system under uncertainty in the State of
of the literature and framework for future investigation. Georgia: a case study. IEEE Transactions on Electronics
Journal of Business Logistics, 19 (1), 85–102. Packaging Manufacturing, 29, 150–162.
Chang, N.-B. and Wei, Y.L., 2000. Siting recycling drop-off Klausner, M. and Hendrickson, C.T., 2000. Reverse-logistics
stations in urban area by genetic algorithm-based fuzzy strategy for product take-back. Interfaces, 30 (3), 156–165.
multiobjective nonlinear integer programming modeling. Kleineidam, U., et al., 2000. Optimising product recycling
Fuzzy Sets and Systems, 114, 133–149. chains by control theory. International Journal of
De Brito, M.P. and Dekker, R., 2003a. A framework for Production Economics, 66, 185–195.
Reverse logistics (January). Econometric Institute Report Krikke, H.R., van Harten, A. and Schuur, P.C., 1999a.
Series ERS-2003-045-LIS, Erasmus University Rotterdam. Business case Océ: reverse logistic network re-design for
De Brito, M.P. and Dekker, R., 2003b. Modelling product copiers. OR Spektrum, 21, 381–409.
returns in inventory control – exploring the validity of Krikke, H.R., van Harten, A. and Schuur, P.C., 1999b.
general assumptions. Production Economics, 81–82, 225–241. Business case Roteb: recovery strategies for monitors.
De Brito, M.P., Dekker, R. and Flapper, S.D.P., 2003. Computers and Industrial Engineering, 36, 739–757.
Reverse logistics – a review of case studies (April). ERIM Kroon, L. and Vrijens, G., 1995. Returnable containers: an
Report Series ERS-2003-012-LIS, Erasmus University example of reverse logistics. International Journal of Physical
Rotterdam. Distribution & Logistics Management, 25 (2), 56–68.
Del Castillo, E. and Cochran, J.K., 1996. Optimal short Krumwiede, D.W. and Sheu, C., 2002. A model for reverse
horizon distribution operations in reusable container logistics entry by third-party providers. Omega, 30, 325–333.
systems. Journal of the Operational Research Society, 47, Linton, J. and Johnston, D.A., 1999. A decision support
48–60. system for planning remanufacturing at Nortel
Dı́az, A. and Fu, M.C., 1997. Models for multi-echelon Networks. Interfaces, 30 (6), 17–31.
repairable item inventory systems with limited repair Listeş, O. and Dekker, R., 2005. A stochastic approach to a
capacity. European Journal of Operational Research, 97, case study for product recovery network design.
480–492. European Journal of Operational Research, 160, 268–287.
Duhaime, R., Riopel, D. and Langevin, A., 2001. Value Louwers, D., et al., 1999. A facility location allocation
analysis and optimization of reusable containers at model for reusing carpet materials. Computers and
Canada post. Interfaces, 30 (3), 3–15. Industrial Engineering, 36, 855–869.
Farrow, P.H., Johnson, R.R. and Larson, A.L., 2000. Maslennikova, I. and Foley, D., 2000. Xerox’s approach to
Entrepreneurship, innovation, and sustainability strate- sustainability. Interfaces, 30 (3), 226–233.
gies at Walden Paddlers, Inc. Interfaces, 30 (3), 215–225. McGavis, D., 1994. The energy bucket and a not-so-drop-
Flapper, S.D.P., 1996. Logistic aspects of reuse. In: First in-the-bucket portion of the waste stream, consumables.
International Working Seminar on Reuse, 11–13 In: Proceedings, IEEE International Symposium on
November 1996, Eindhoven University of Technology Electronics and the Environment, 2–4 May 1994, San
Eindhoven, The Netherlands, 109–118. Francisco, CA. 2267–2272.
 260 T. J. Barker and Z. B. Zabinsky

McIntyre, K., 2007. Delivering sustainability through Seitz, M.A. and Peattie, K., 2004. Meeting the closed-loop
supply chain management. In: D. Waters, ed. Global challenge: the case of remanufacturing. California
logistics: new directions in supply chain management. Management Review, 46 (2), 74–89.
London: Kogan Page Limited, 238–252. Spengler, T., et al., 1997. Environmental integrated
Meade, L., Sarkis, J. and Presley, A., 2007. The theory and production and recycling management. European
practice of reverse logistics. International Journal of Journal of Operational Research, 97, 308–326.
Logistics Systems and Management, 3 (1), 56–84. Staikos, T. and Rahimifard, S., 2007. A decision-making
Meyer, H., 1999. Many happy returns. Journal of Business model for waste management in the footwear industry.
Strategy, 20 (4), 27–31. International Journal of Production Research, 45 (18–19),
Nagel, C. and Meyer, P., 1999. Caught between ecology 4403–4422.
and economy: end-of-life aspects of environmentally Thierry, M., et al., 1995. Strategic issues in product
conscious manufacturing. Computers and Industrial recovery management. California Management Review,
Engineering, 36, 781–792. 37 (2), 114–135.
Realff, M.J., Ammons, J.C. and Newton, D., 2000. Thomas, Jr., M., 1997. Emerging technology: production
Strategic design of reverse production systems. scheduling matures. IIE Solutions, January, 25–29.
Computers and Chemical Engineering, 24, 991–996. Toktay, L.B., Wein, L.M. and Zenios, S.A., 2000.
Realff, M.J., Ammons, J.C. and Newton, D., 2004. Robust Inventory management of remanufacturable products.
Downloaded by [Australian National University] at 19:18 07 January 2015

reverse production system design for carpet recycling. IIE Management Science, 46 (11), 1412–1426.
Transactions, 36, 767–776. Wang, C.-H., Even, Jr., J.C. and Adams, S.K., 1995. A
Rubio, S., Chamorro, A. and Miranda, F.J., 2008. mixed-integer linear model for optimal processing and
Characteristics of the research on Reverse Logistics transport of secondary materials. Resources, Conservation
(1995–2005). International Journal of Production and Recycling, 15, 65–78.
Research, 46 (4), 1099–1120. Yender, G.L., 1998. Battery recycling technology and
Rudi, N., Pyke, D.F. and Sporsheim, P.O., 2000. Product collection processes. In: Proceedings, IEEE international
recovery at the Norwegian National Insurance symposium on electronics and the environment, 4–6 May
Administration. Interfaces, 30 (3), 166–179. 1998, Oak Brook, IL, 30–35.