A multi-objective hybrid evolutionary approach for buffer allocation in open serial production lines

• Published in: Journal of intelligent manufacturing Vol. 31; no. 1; pp. 33 - 51
• Main Authors: Yelkenci Kose, Simge, Kilincci, Ozcan
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.01.2020; Springer Nature B.V
• Subjects: Advanced manufacturing technologies; Algorithms; Assembly lines; Buffer inventories; Buffers; Business and Management; Classification; Comparative studies; Computer simulation; Control; Discrete event systems; Evolutionary algorithms; Genetic algorithms; Machines; Manufacturing; Maximization; Mechatronics; Multiple objective analysis; Operations management; Optimization; Pareto optimum; Processes; Production; Production lines; Production planning; Robotics; Simulated annealing; Simulation; Sorting algorithms; Hybrid meta-heuristics; Multi-objective optimization; Production lines; Buffer allocation problem
• ISSN: 0956-5515; 1572-8145
• DOI: 10.1007/s10845-018-1435-6

The buffer allocation problem is of particular interest for operations management since buffers have a considerable impact on capacity improvement in production systems. In this study, the buffer allocation is solved to optimize two conflicting objectives of maximizing the average system production rate and minimizing total buffer size. A hybrid evolutionary algorithm-based simulation optimization approach is proposed for the multi-objective buffer allocation problem (MOBAP) in open serial production lines. As a search methodology, the Pareto optimal set is derived by hybrid approach using elitist non-dominated sorting genetic algorithm (NSGA-II) and a special version of a multi-objective simulated annealing. As an evaluative tool, discrete event simulation modeling is used to estimate the performance measures for the production systems. To demonstrate the efficacy of the proposed hybrid approach, a comparative study is provided for the MOBAP in various serial line configurations. The comparative results show that the hybrid method has a considerable potential to minimize the total buffer space by appropriately allocating space to each buffer while maximizing average production rate.