关键词: 双边装配线平衡, 离散人工蜂群算法, 变邻域搜索, 工位选择策略, 操作选择策略

Abstract: Aiming at the discreteness and sequence correlation of type-I two-sided assembly line balancing,an improved Discrete Artificial Bee Colony (DABC) algorithm was proposed.In this algorithm,new heuristic objective was added as new nectar,the employed bees was arranged to make crossover operator to expand search space,and the onlooker bees was utilized to do variable neighbor search to enhance local search.In addition,an improved station-task decoding method was put forward to reduce idle times due to direction and precedence constraints.In this decoding method,the station selection strategy chose the side with larger capacity to balance the workload on the two sides of a current mated-station,while task selection strategy selected the tasks which didnt result in idle time at first so as to reduce sequence-dependent idle time.The comparison among three different decoding schemes proved that the proposed decoding scheme could reduce idle time effectively.By comparing with seven recent meta-heuristics,DABC algorithm found all the current best solutions and two new optimal solutions.

Key words: two-sided assembly line balancing, discrete artificial bee colony algorithm, variable neighbor search, station selection strategy, task selection strategy