Evolutionary Algorithms for Multi-Criteria Supply Chain Optimization

Evolutionary algorithms based on multi-objective optimization provide a versatile platform in the decision-making process of complex supply chains and provide groups of trade-off solutions at the same time. Some of the key MOEAs such as the entry-level Genetic Algorithm through to advanced ones such as NSGA-II/III, SPEA2, MOEA/D, IBEA, and PAES have been surveyed and their technical methods and selection processes discussed. The strengths of each algorithm are as follows: NSGA-II is characterized by elitist sorting, SPEA2 is characterized by an archive scheme, MOEA/D is characterized by decomposition, IBEA by indicator metrics, PAES by simplicity, and NSGA-III by a large number of objectives. The weaknesses differ between the cost of computation and sensitivity of the parameter. The case studies in supply chain entail empirical comparisons which reveal that there is no single approach that prevails. Supply chain optimization trends are moving towards hybrid (MOEAs with simulation or local search), many-objective, and domain-specific optimizations (e.g., fuzzy criteria). All in all, MOEAs have continued to be an effective instrument in supply chain multi-criteria decision problems. The choice of the right algorithm varies with the number of goals, the desired Pareto dispersion, and the computational capabilities. Future development can focus on automated algorithm configuration and interactive MOEAs that update preferences of the decision-maker in real time.

[1]            K. Deb, A. Pratap, S. Agrawal, and T. Meyarivan, “A fast and elitist multi-objective genetic algorithm: NSGA-II,” IEEE Trans. Evol. Comput., vol. 6, no. 2, pp. 182–197, Apr. 2002.   [2]            E. Zitzler, M. Laumanns, and L. Thiele, “SPEA2: Improving the strength Pareto evolutionary algorithm,” in Proc. EUROGEN (Lecture Notes in Comput. Sci. vol. 1143), 2001, pp. 95–100.   [3]            J. D. Knowles and D. W. Corne, “The Pareto Archived Evolution Strategy: A new baseline multiobjective optimization algorithm,” in Proc. IEEE CEC, 1999, pp. 98–105.   [4]            E. Zitzler and S. Künzli, “Indicator-based selection in multiobjective search,” in Proc. PPSN, 2004, pp. 832–842.   [5]            Q. Zhang and H. Li, “MOEA/D: A multiobjective evolutionary algorithm based on decomposition,” IEEE Trans. Evol. Comput., vol. 11, no. 6, pp. 712–731, Dec. 2007.   [6]            S. Türk, E. Özcan, and R. John, “Multi-objective optimization in inventory planning with supplier selection,” Expert Syst. Appl., vol. 67, pp. 115–124, 2017.   [7]            A. Brintrup and A. Puchkova, “Multi-objective optimisation of reliable product-plant network configuration,” Appl. Netw. Sci., vol. 3, no. 1, 2018.   [8]            L. Amodeo, H. Chen, and A. El Hadji, “Multi-objective supply chain optimization: An industrial case study,” in Applications of Evolutionary Computing (LNCS vol. 4448), 2007, pp. 732–741. [9]            K. Deb and H. Jain, “An evolutionary many-objective optimization algorithm using reference-point-based nondominated sorting approach, Part I: Solving problems with box constraints,” IEEE Trans. Evol. Comput., vol. 18, no. 4, pp. 577–601, Aug. 2014.