Abstract: In view of complexity of disassembly line balancing problems, an improved multi-objective artificial fish swarm algorithm was proposed based on Pareto set. In order to improve the searching ability of artificial fish, a random crossover operation of genetic algorithm was introduced to guide the artificial fish to the direction of global optimal disassembly directions. The crowding distance was adopted to filter the non-inferior solutions in the prey, swarm and follow processes of the artificial fish to realize the diversity of each behavior solution results. By employing the strategy of elite reservation, the non-inferior solutions in external file were added to the next iteration of the algorithm, which speeded up the convergence rate of the algorithm. Finally, the proposed algorithm was applied to the disassembly instances of different scales, and the results confirm the effectiveness and superiority of the proposed algorithm by comparing with the existing algorithms. 

Key words: disassembly line balancing problem, multi-objective optimization, Pareto set, artificial fish swarm algorithm

摘要： 针对拆卸线平衡问题的复杂性，提出了一种改进的基于Pareto解集的多目标人工鱼群算法进行求解。为提高人工鱼觅食时的寻优能力，引入遗传算法的随机交叉操作，指导人工鱼向全局最优拆卸方向觅食。通过拥挤距离不断筛选人工鱼觅食、聚群和追尾过程中的非劣解，实现了各行为结果的多样性。采用精英保留策略，将外部档案中的非劣解添加到算法下次迭代的种群中，加快了算法的收敛。通过对不同规模的拆卸实例进行求解，并将其与已有算法进行对比，验证了所提算法的有效性和优越性。

关键词: 拆卸线平衡问题, 多目标优化, Pareto解集, 人工鱼群算法