Abstract: The integrated optimization problems for pull production systems composed of one mixed-model assembly line and one part fabrication line with identical parallel machines and limited intermediate buffers are investigated. Two objectives are considered simultaneously：Minimizing the total variation in parts consumption in the assembly line and minimizing the makespan in the fabrication line. The integrated optimization framework, the mathematical model for the assembly line and a procedure to construct a complete schedule for the fabrication line are presented. A multi-objective genetic algorithm (MOGA) is proposed for solving the problem, in which a three-stage real number encoding method is put forward, the Pareto ranking method and the sharing function method are employed to evaluate the individuals’ fitness, the selection, crossover, mutation operators and the elitist strategy are designed, which guarantees the dispersity and uniformity of the solutions. The feasibility and efficiency of the MOGA is shown by comparison with a multi-objective simulated annealing algorithm (MOSA). The computational results show that satisfactory non-dominated solution set can be obtained by the MOGA.

Key words: Integrated optimization, Mixed-model fabrication/assembly system, Multi-objective genetic algorithm, Multi-objective simulated annealing algorithm