Modelling and optimization of material purchasing process in apparel supply chain

Abstract

Facing the increasingly fierce competition and the fast-changing customer demand, apparel companies have striven to lower purchasing costs and shorten purchasing lead time using systematic and effective methods of material purchasing decision-making. The purpose of this research is to develop intelligent algorithm-based methodologies for material purchasing decision-making of the apparel supply chain. An effective supplier selection and order allocation (SSOA) model for material purchasing decision-making in the apparel supply chain is developed through integrating three types of apparel material purchasing problem, namely 1) supplier evaluation and ranking problems at the purchasing pre-selection stage, 2) supplier selection and order allocation for single-item purchasing problems, and 3) supplier selection and order allocation for multi-item purchasing problems at the final purchasing selection stage. On the basis of fuzzy extent analytic hierarchy processes (FEAHP), dynamic programming (DP) and improved differential evolution (DE), these three types of problem are formulated mathematically and solved by effective methodologies. Supplier pre-evaluation and ranking at the purchasing preparation stage is a multiple criteria decision problem involving qualitative and quantitative factors. In this stage, the decision maker needs to identify critical decision criteria and then evaluate, rank and preselect potential suppliers with respect to those criteria. With consideration for the fuzziness of data involved in deciding the preferences of multiple decision variables, the fuzzy-extended analytic hierarchy process (FEAHP) -based methodology is developed to determine the multiple decision criteria. In the final stage of material purchasing, this research investigates common material (e.g. white fabric) purchasing (i.e. single-item multiple-period purchasing). An integrated approach, including FEAHP, multi-objective programming (MOP) and dynamic programming (DP), is developed to identify ultimate suppliers and determine optimum order quantities among selected suppliers to minimize material purchasing risks and total material purchasing costs, with consideration for various types of customer demand, supplier capacity and material prices given by suppliers. In dealing with fashion accessories purchasing in the final stage of material purchasing, multi-item purchasing, one-time purchasing, price discount and on-time delivery are considered. An improved differential evolution (DE) algorithm and a probability theory-based optimization model are developed to solve a stochastic discrete multi-objective problem. In this model, uncertain delivery delay and uncertain product discount are determined using the probability theory. The ultimate number of selected suppliers and the optimal order allocation strategy are given by an improved DE algorithm, namely the composite discrete differential evolution (CoDDE) algorithm. Extensive experiments based on industrial data are conducted to validate the proposed models and evaluate the performance of the proposed methodologies. The experiment results demonstrate the effectiveness of the proposed model and methodologies for material purchasing decision-making of the apparel supply chain.