A Reverse Logistics Provider Selection Framework
[004-0351]
Tuğba Fırdolaş, Department of Industrial Engineering, Yildiz Technical University, Yildiz, Istanbul, Turkey
Semih Önüt, Department of Industrial Engineering, Yildiz Technical University, Yildiz, Istanbul, Turkey
Elif Kongar[1], Center for Industrial Ecology, Yale University, New Haven CT, USA.
ABSTRACT
Today, sustainable development has become the central focus of many researchers and companies. To achieve the predetermined goals and objectives, a company must be able to respond to increasing consumer demand for “green” products and to implement environmentally conscious plans. Reverse logistics plays a crucial role in achieving this since it enables customers with the opportunity to return the warranted and/or defective products. Hence, an efficient reverse logistics structure provides a competitive advantage to companies as well as a significant return on investment. In this regard, the selection of third-party providers, being one of the most important issues in the overall reverse logistics structure, is a crucial step in reverse logistics related practices.
This study aims to assist the decision makers in determining the “most appropriate” third-party logistics provider via an alternative approach based on an integrated model using artificial neural networks and fuzzy logic. A case study is also included to demonstrate the logical steps of the model.
Keywords: Reverse logistics, artificial neural networks, fuzzy AHP, vendor selection.
1. Introduction
Today supply chain has become a key component in building and maintaining an organization’s competitiveness. One of the core elements of effective supply chain management is logistics. Logistics is not only delivering of goods but also offering the opportunity for stock to be returned to suppliers via a feedback loop. Along with the need or potential for the re-use or recycling of unwanted stock, increasing concerns over environmental responsibility and increasing opportunities for cost savings or revenues from returned products prompted some researchers to formulate a new concept labeled as “reverse logistics”. Traditionally, reverse logistics is an activity within organizations delegated to the customers with warranted or defective products would return them to their suppliers (Meade and Sarkis, 2002).
Handling the returns can be considered as a problem when the following symptoms occur: the arrival rate of returns is larger than the processing or disposal of the items, large amount of returns inventory held in the warehouse taking up space and other storage costs, unidentified or unauthorized returns, lengthy cycle processing times, unknown total cost of the returns process, and loss of customer confidence in the repair activity (Schwartz, 2000). In this respect, reverse logistics suppliers play an important role in helping organizations to complete the loop for products offered by those organizations. This study aims to assist managers in determining which third-party logistics provider to collaborate in the reverse logistics process with an alternative approach based on an integrated conceptual framework using artificial neural networks and fuzzy logic. Literature review is presented in Section 2, and the nomenclature is provided in Section 3. Section 4 provides a brief introduction to the evaluation approaches for the integrated conceptual framework followed by a case study described in Section 5. Section 6 concludes the study and its findings.
2. LITERATURE REVIEW
2.1. Reverse Logistics and Reverse Supply Chain
Environmental factors are rapidly emerging as an important issue for business and management to consider. To achieve business goals and objectives, a company must reply to increasing consumer demand for and implement environmentally responsible plans. Along with the forward supply chain, the organization needs to consider the impact of reverse logistics. A reverse logistics defines a supply chain that is redesigned to efficiently manage the flow of products or parts destined for remanufacturing, recycling, or disposal and to effectively utilize resources (Dowlatshahi, 2000). Emergence of reverse logistics enables to provide a competitive advantage and significant return on investment with an indirect effect on profitability.
Reverse logistics is a relatively new topic in the academic research area. Few analytical models exist which assist in reverse logistics strategic decisions. Past literature has focused on a variety of issues in reverse logistics including production planning and inventory control in remanufacturing, facility location decisions, resource allocation and flows. Fleischmann et al.(1997) surveyed the recently emerged field of reverse logistics, addressed the implications of the emerging reuse efforts and reviewed the mathematical models proposed in the literature. 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. Spengler et al.(1997) developed a mixed-integer linear programming model for recycling of industrial by products which is applied to the German steel industry. Sarkis (1998) proposed a model that considers the systemic and hierarchical relationships among a numbers of decision and environmental factors facing organizations. Carter and Ellram (1998) reviewed the literature on reverse logistics and suggested some critical factors in the reverse logistics process. Specifically, they proposed that sincere shareholder commitment and top management support are necessary for the continued success of a reverse logistics program. Sarkis (1999) suggested a two-stage methodology that integrates managerial preferences to help evaluate quantitative data for the purpose of selecting an environmentally conscious manufacturing (ECM) program. 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 this area. Knemeyer et al.(2002) proposed a qualitative methodology to examine the feasibility of designing a reverse logistics system to recycle and/or refurbish end-of-life computer that were deemed no longer useful by their owners. Krumwiede and Sheu (2002) reviewed current industry practices in reverse logistics, examined the issues and processes that had to be addressed to engage in the reverse logistics business. De Brito et al. (2002) gave an overview of scientific literature that described and discussed cases of reverse logistics activities in practice by considering over sixty case studies. Tibben-Lembke and Rogers (2002) compared and contrasted forward and reverse logistics in a retail environment, with the focus on the reverse flow of product. Kongar and Gupta (2002) used multi-criteria decision-making technique to study disassembly-to-order systems. Sarkis (2003) focused on the components and elements of green supply chain management as a foundation for the decision framework with the applicability of a dynamic non-linear multiattribute decision model. Chouinard et al. (2005) dealt with problems related to the integration of reverse logistics activities within an organization and to the coordination of this new system. Ravi et al. (2005) proposed a combination of balanced scorecard and analytic network process to provide a more realistic and accurate representation for conducting reverse logistics operations for end-of-life computers.
Studies in the literature mainly focus and evaluate the supply chain from a single perspective. Published work includes the effects of basic supply chain components such as, inventory management, information technologies, demand forecast, quality management, and logistics on the supply chain performance, as well as the effects of more specific issues such as, uncertainty and complexity, postponement, product and market variability (Kongar, 2005).
Many organizations are hiring third-party logistics providers to implement reverse logistics programs designed to retain value by getting products back. Suppliers play a crucial role in supply chain and hence in the long term viability of a company. In this regard, the selection of third-party logistics provider issue is increasingly becoming an area of reverse logistics concept and practice. Selecting the right suppliers significantly reduces the purchasing cost and improves corporate competitiveness. There are various techniques for selecting suppliers/partners such as matrix or weight approaches, mathematical programming approaches (i.e., linear programming, mixed integer-programming, goal programming, multi-objective programming and non-linear programming, multi-criteria decisions methods) (Ghodsypour and O’Brien (1998), Liu et al.(2000), Dulmin and Mininno(2003), Talluri and Narasimhan (2003), probabilistic methods, artificial intelligent techniques (i.e., genetic algorithms, neural networks, fuzzy logic) (Wei et al.1997), Ding et al.(2003), Choy et al.(2003)) and some integrated approaches (Kumar et al. (2004)).
As listed above, even though there exists a wide range of studies regarding the selection of forward logistics providers; only a few examined the selection of reverse logistics providers. In the respect, Meade and Sarkis (2002) proposed a strategic model that links with operational characteristics in prepared according to the selection of third-party logistics providers by using the analytical network process technique. In their study, basically four decision factors were used to facilitate the selection of a third party reverse logistics provider, respectively; product lifecyle position, organizational performance criteria, reverse logistics process functions, organizational role of reverse logistics.
Selecting reverse logistics vendors from a large number of possible suppliers with various levels of capabilities and potential is a difficult and time-consuming task that is often driven by multiple criteria. Hence, this study has been prepared on this motivation considering the selection methodologies and criteria of third party logistics provider to partner in reverse logistics process. Performance criteria that affect the selection process of reverse logistics providers are as follows.
2.2. Performance Criteria Selection
Supplier selection decisions are complicated by the fact that various criteria must be considered through out the decision making process. The analysis of such criteria and measuring the performances of suppliers have been the focus of many researchers and purchasing practitioners for approximately four decades.
A firm must examine key strategic factors consisting of: (a) strategic costs, (b) overall quality, (c) customer service, (d) environmental concerns and (e) legislative concerns. Also key operational factors needing to be addressed consist of cost-benefit analysis, transportation, warehousing, supply management, remanufacturing and recycling, and packaging (Knemeyer, 2002). Persson and Olhager (2002) proposed a supply chain simulation study to evaluate alternative supply chain designs with respect to quality, lead-times and costs and to increase the understanding of the interrelationships among these and other parameters, relevant for the design and operations of a supply chain.
Most performance criteria are defined in literature proposing strategic performance measurement system structure and appropriate performance metrics to be used in the measurement process. Kongar (2005) classified the performance metrics into five perspectives as financial perspective, customer perspective, internal business processes perspective (pre-production, production, post-production), environmental perspective and learning and growth perspective. This classification has been used in the proposed methodology to evaluate the reverse logistics providers. It helped to simplify and to standardize the vendors’ selection process by deciding the appropriate performance criteria. For more information on the performance criteria see Kongar (2005).
Firms must analyze and document the significance of several of factors, converting instinctive qualitative indicators to concise empirical measures. The proposed model has a two-stage methodology that considers tangible, intangible, quantitative, qualitative, strategic and operational factors in the decision. The performance criteria that behave as the providers’ selection criteria for this methodology have been chosen from Kongar’s (2005) classification list.
3. NOTATION
The nomenclature used in this paper is as follows:
TFN / Triangular fuzzy number/ Membership function of M TFN
/ nxn fuzzy matrix
/ Fuzzy numbers
/ M TFN
/ Fuzzy synthetic extent value
d / Intersection point between two fuzzy numbers
/ Weight vector
/ Nonfuzzy number
xp / Input signals for ANN
wkp / Synaptic weights of neuron k for ANN
uk / Linear combiner output
θk / Threshold for ANN
Φ(.) / Activation function
yk / Output signal
T / Desired output
O / Actual output
t / Period
p / Sample number
PDR / Perfect delivery ratio
ROD / On time delivery ratio
UOP / Unit operation cost
CUR / Capacity usage ratio
MSI / Increment in market share
RD / Research and Development ratio
CEN / Environmental expenditures
CLI / Customer loyalty index
CSI / Customer satisfaction index
4. AN EVALUATION APPROACH FOR PROVIDER SELECTION
Fuzzy logic and artificial neural networks both being computational intelligence technologies, it is easily justifiable that these two are combined to produce synergetic effects through an integrated frame approach which takes the advantage of benefits and at the same time counterbalances the drawbacks of these two technologies. With this motivation, proposed integrated conceptual framework includes both of these pre-mentioned artificial intelligence techniques.
4.1. Fuzzy Analytic Hierarchy Process
4.1.1. Structure
The analytic hierarchy process (AHP) suggested by Saaty (1988), is one of the extensively used multi-criteria decision-making methods. It is easier to understand and it can effectively handle both qualitative and quantitative data. AHP solution approach’s steps are (Saaty, 1994); (1) Problem decomposition – the problem is decomposed into elements (which are grouped on different levels to form a chain of hierarchy) and each element is further decomposed into sub-elements until the lowest level of the hierarchy (Figure 1). (2) Comparative analysis – the relative importance of each element at a particular level will be measured by a procedure of pairwise comparison. The decision makers provide numerical values for the priority of each element using a rating scale. (3) Synthesis of priorities – the priority weights of elements at each level will be computed using eigenvector or least square analysis. The process is repeated for each level of the hierarchy until a decision is finally reached by overall composite weights.
The AHP approach requires humans to translate their perceptions into numerical scales, frequently through mechanisms like a Likert scale (1-3-5, 1-3-9 or 1-5-9) to quantify the ‘‘decision makers’ strength of feeling between any two attributes’’ with respect to any given criterion. The process makes use of a suitable process to estimate relative weights of the decision elements (such as eigenvalues’) and culminates into their aggregation in order to arrive at the outcome. However human assessment on the relative importance of individual customer requirements is always subjective and imprecise. The linguistic terms that people use to express their feelings or judgment are vague (Büyüközkan et al., 2004).