A Hybrid Approach using the Bees Algorithm and Fuzzy-AHP for Supplier Selection

Introduction

Suppliers are one of the most important elements in a supply chain. A supply chain will be more robust and sustainable with stronger suppliers. In order to have robust suppliers in the chain, a strong evaluation process needs to be considered, as the supplier is the provider of raw materials/parts for main manufacturer, which needs to be supplied on time with right amount (Choi and Hartley, 1996). Thus the selection of appropriate suppliers directly/indirectly affects the performance of the entire supply chain. Supplier selections process has been widely modelled in literature (Lambiase et al, 2013). Several qualitative factors such as quality, cost, service level, supplier profile, risk, and quantitative factors such as capacity and demand should be considered during the supplier selection process (Ghodsypour and O'brien, 1998). The process requires an intense and strong decision making stage to determine the right suppliers and order quantities. To determine the correct decision under these multi criteria, a decision making process has to be utilized. In literature, several Multi Criteria Decision-Making (MCDM) methods have been proposed (Ho et al, 2010). Analytical Hierarchy Process (AHP) is one of the most common methods used for supplier selection since allows to make a decision through a pair-wise comparison of the criteria involved in the decision process (Saaty, 1988). However, in order to better support the uncertainty in the human judgment, fuzzy sets theory (Zadeh, 1965) has been added to the conventional AHP by replacing fixed value judgment with fuzzy intervals. In other words, in the Fuzzy-AHP (Kahraman et al, 2003) the pair-wise comparison of each criterion over another is not performed by crisp numbers, but by means of membership functions. Furthermore, in order to consider quantitative parameters and constraints, an optimization algorithm has to be utilized in order to allocate the optimum amount of purchases from different suppliers. In literature, several optimization algorithms have been applied to solve the supplier selection problem. Among these, a recent promising swarm-based meta-heuristic optimization algorithm is the Bees Algorithm (BA) (Yuce et al, 2013). The BA has been proposed by Pham et al. in 2005 (Pham et al, 2005) and since then has been used to solve different types of optimization problems as supply chain design (Mastrocinque et al, 2013), job shop scheduling (Packianather et al, 2014). The algorithm has a strong ability to perform better in most cases compare to other well-known optimization algorithms (Pham and Castellani, 2009). The algorithm can utilize both global search and local search during the search process, which are the main advantages compare to other global search based algorithms such as Genetic Algorithm (Davis, 1991). The algorithm utilizes the fitness evaluation operation to find the global optimum solutions, which provides another advantage to the algorithm compared to other bee based algorithms such as Artificial Bee Colony (ABC) (Karaboga, 2005). In this chapter, the Bees Algorithm has been used to determine the optimum amount of raw material units from each supplier under several constraints.

The main objectives of this chapter are firstly to give an overview of the supplier selection problem, then to utilize a hybrid approach using the Bees Algorithm and Fuzzy-AHP to determine the optimum purchase quantity. The authors want to provide an alternative methodology to decision makers for supplier selection problem.