Multi-objective biogeography-based optimization for supply chain network design under uncertainty

• Published in: Computers & industrial engineering Vol. 85; pp. 145 - 156
• Main Authors: Yang, Guo-Qing, Liu, Yan-Kui, Yang, Kai
• Format: Journal Article
• Language: English
• Published: New York Elsevier Ltd 01.07.2015; Pergamon Press Inc
• Subjects: Algorithms; Approximation; Approximation approach; Biogeography; Biogeography based optimization; Demand; Mathematical models; Mathematics; Molybdenum; Multi-objective optimization; Networks; Operations research; Optimization; Optimization algorithms; Studies; Supply chain management; Supply chain network design; Supply chains; Weighted mean deviation; Biogeography based optimization; Multi-objective optimization; Supply chain network design; Weighted mean deviation; Approximation approach
• ISSN: 0360-8352; 1879-0550
• DOI: 10.1016/j.cie.2015.03.008

•We develop a new two-stage optimization method for MO-SCND under uncertainty.•The proposed model includes uncertain transportation costs and customer demands.•An approximation method is adapted to approximate original optimization model.•A new MO-BBO algorithm is designed to solve the approximate optimization model.•Computational results demonstrate the effectiveness of the MO-BBO algorithm. This paper proposes a new two-stage optimization method for multi-objective supply chain network design (MO-SCND) problem with uncertain transportation costs and uncertain customer demands. On the basis of risk-neutral and risk-averse criteria, we develop two objectives for our SCND problem. We introduce two solution concepts for the proposed MO-SCND problem, and use them to define the multi-objective value of fuzzy solution (MOVFS). The value of the MOVFS measures the importance of uncertainties included in the model, and helps us to understand the necessity of solving the two-stage multi-objective optimization model. When the uncertain transportation costs and customer demands have joined continuous possibility distributions, we employ an approximation approach (AA) to compute the values of two objective functions. Using the AA, the original optimization problem becomes an approximating mixed-integer multi-objective programming model. To solve the hard approximating optimization problem, we design an improved multi-objective biogeography-based optimization (MO-BBO) algorithm integrated with LINGO software. We also compare the improved MO-BBO algorithm with the multi-objective genetic algorithm (MO-GA). Finally, a realistic dairy company example is provided to demonstrate that the improved MO-BBO algorithm achieves the better performance than MO-GA in terms of solution quality.