Solving a Real-world Supply-Chain Management Problem Using a Bilevel Evolutionary Approach COIN Report Number 2017009

Supply-chain management problems are common to most industries and they involve a hierarchy of subtasks, which must be coordinated well to arrive at an overall optimal solution. Such problems also involve more than one stake-holders, such as the supplier, the transport companies that supply goods from source to destinations, and multiple management levels within the supplier. Thus, these problems involve a hierarchy of decision-makers, each having its own objectives and constraints, but importantly requiring a coordination of their actions to make the overall supply chain process optimal from cost and quality considerations. In this paper, we consider a specific supply-chain management problem from an industry, which involves two levels of coordination: (i) yearly strategic planning in which a decision on establishing an association of every destination point with a supply point must be made so as to minimize the yearly transportation cost, and (ii) weekly operational planning in which, given the association between a supply and a destination point, a decision on the preference of available transport carriers must be made for multiple objectives: minimization of transport cost and maximization of service quality and satisfaction of demand at each destination point. Thus, the resulting problem is bilevel, for which the operational planning problem is nested within the strategic planning problem. The problem is also multi-objective in nature. Moreover, the problem involves several practical challenges, such as uncertainty in demand at each destination, non-linearity and non-differentiability of the cost model, large dimensions, and others. We propose a customized multiobjective bilevel evolutionary algorithm, which is computationally tractable. We then present results on state-level and ZIP-level accuracy (involving about 40,000 upper level variables) of destination points over the mainland USA. We compare our proposed method with current non-optimization based practices and report a considerable cost saving.

[1]  Kalyanmoy Deb,et al.  Derived heuristics-based consistent optimization of material flow in a gold processing plant , 2018 .

[2]  Kalyanmoy Deb,et al.  Breaking the Billion-Variable Barrier in Real-World Optimization Using a Customized Evolutionary Algorithm , 2016, GECCO.

[3]  Xiaofan Lai,et al.  A multi-objective optimization for green supply chain network design , 2011, Decis. Support Syst..

[4]  Gao Zi-You,et al.  A bi-level programming model and solution algorithm for the location of logistics distribution centers , 2008 .

[5]  Takeshi Uno,et al.  An evolutionary multi-agent based search method for stackelberg solutions of bilevel facility location problems , 2008 .

[6]  Shi Feng Bi-level Simulated Annealing Algorithm for Facility Location , 2007 .

[7]  Turan Paksoy,et al.  A genetic algorithm approach for multi-objective optimization of supply chain networks , 2006, Comput. Ind. Eng..

[8]  Xiaoling Sun,et al.  Nonlinear Integer Programming , 2006 .

[9]  Kalyanmoy Deb,et al.  A fast and elitist multiobjective genetic algorithm: NSGA-II , 2002, IEEE Trans. Evol. Comput..

[10]  Martine Labbé,et al.  A Bilevel Model for Toll Optimization on a Multicommodity Transportation Network , 2000, Transp. Sci..

[11]  D. Lambert,et al.  Supply Chain Management: Implementation Issues and Research Opportunities , 1998 .

[12]  Athanasios Migdalas,et al.  Bilevel programming in traffic planning: Models, methods and challenge , 1995, J. Glob. Optim..

[13]  Michael Florian,et al.  Optimizing frequencies in a transit network: a nonlinear bi-level programming approach , 1995 .

[14]  Jerome Bracken,et al.  Mathematical Programs with Optimization Problems in the Constraints , 1973, Oper. Res..

[15]  Heinrich von Stackelberg theory of the market economy , 1952 .