针对传统方法求解多目标U型拆卸线平衡问题的不足,提出了一种基于Pareto解集的多目标蚁群遗传算法.在构造初始解阶段,以协同考虑最大作业时间、最小拆卸成本差作为蚂蚁的启发式信息;通过蚁群算法搜索可行拆卸序列,并根据多目标之间的支配关系得到Pareto解集;将蚁群算法的Pareto非劣解作为遗传操作的个体,迸而将遗传操作的结果正反馈于最优拆卸路径上信息素的积累,并采用拥挤距离作为蚂蚁全局信息素更新策略,可以平衡多目标对信息素的影响,使算法快速获得较优解.将所提算法应用于52项拆卸任务算例和某打印机拆卸线实例,在算例验证中,通过对比Pareto蚁群算法,所提算法求得的8个非劣解在3个评价指标上性能分别提高了50.43%、3.25%、14.10%,在实例应用中所提算法求得8种可选平衡方案,从而验证了所提算法的有效性、优越性和实用性.%Owing to the incapability of traditional methods in solving multi-objective U-shaped disassembly line balancing problem(UDLBP),a multi-objective ant colony genetic algorithm based on Pareto set is proposed to solve the UDLBP. In constructing the initial solution phase,the maximum operation time and the minimum disassembly cost difference were collaboratively considered as the heuristic information of the ants. The feasible disassembly sequence was searched using the ant colony algorithm,and the Pareto solution set was obtained based on the dominance relationship among the multiple objectives. Further,the Pareto non-inferior solutions of the ant colony algorithm were used as the chromosomes of the genetic operator. Moreover,the results of the genetic operator were fed back to the accumulation of pheromone on the optimal disassembly sequence. The crowding distance was regarded as the global pheromone update strategy to balance the effect of multi-objective function regarding the pheromone and thereby make the algorithm obtain better solutions readily. The proposed algorithm was applied to 52 disassembly task examples and a printer disassembly line instance. Compared with the Pareto ant colony algorithm,the performances of the three evaluation indicators of the 8 non-inferior solutions obtained using the proposed algorithm are improved by 50.43%,3.25%, and 14. 10%,respectively. In the example application,the proposed algorithm obtained eight balancing schemes,which verifies the effectiveness,superiority,and practicability of the proposed algorithm.
展开▼
