A Multi-objective Particle Swarm Optimization for Assembly Line Design with Station Paralleling

摘要

To solve bi-objective assembly line design problem (ALDP) considering station paralleling and equipment selection, a Pareto-dominance-based method is presented. The two objectives are minimization of investment cost and maximization of availability of the assembly line. A multiobjective particle swarm optimization (MoPSO) is proposed to obtain a set of Pareto solutions through combining the techniques of crowded distance and external Pareto solution archive. The developed solution representation and relating updating mechanism of MoPSO ensures each particle to be a feasible solution. The performance of MoPSO was compared with that of NSGA-II against two cases. The comparison results show the effectiveness of the MoPSO. The computation results also indicate that the MoPSO is superior to the NSGAII for the ALDP with respect to solution quality and computational efficiency.