采用传统的关键邻域搜索方法求解作业车间调度问题时,往往容易陷入局部极值而且难以跳出。为此,提出了一种具有动态调整能力的全局邻域交换策略,该策略有可能产生大量的不可行调度,需要一种筛选方法加以过滤。证明了一个新的邻域交换性质,利用该性质可以对所得调度方案作可行性约束判定,从而有效地过滤掉不可行调度。在此基础上,提出了一种求解作业车间调度问题的算法。最后,取不同规模的Benchmark问题算例对该算法进行测试,结果表明,无论从解的质量还是计算时间都取得了较好的效果。
Traditional critical neighborhood algorithms for Job Shop scheduling problem were easily trapped into local optimal and hardly to escape. To deal with this problem, a global neighborhood swapping strategy with dynamic adapatability was proposed. However, this new strategy might possibly induce infeasible solutions. Thus, a new proposition concerning the neighborhood swapping strategy was presented and proved, which could be used to verify whether a neighborhood swapping was acceptable or not. Based on this global neighborhood strategy, a new algorithm was developed and tested by a group of benchmark instances. The results indicated that the new algorithm obtained satisfactory results both on solutions' quality and computation time.