网络化集成制造模式下调度问题的混合遗传算法

研究了网络化集成制造模式下的JOB-SHOP调度问题，即各任务有具体交货期约束的提前／拖期不同惩罚系数的复杂调度问题；提出了一种新的编码方案和相应的交叉算子，进而提出混合遗传算法；成功解决了上述问题，即找到一个最优调度，使在交货期窗口外完工的任务所受到的提前／拖期惩罚的总和最小．该算法解码复杂度低、通用性强、罗拔性强，最后用算例说明了算法的有效性并指出了工件车间调度问题中尚待解决的问题．

Hybrid Genetic Algorithm for a Scheduling Problem with a Network Integrated Manufacturing System

The job shop scheduling problems with a network integrated manufacturing system, namely different early/tardy penalty coefficients scheduling problems with due date constraints is studied. An ameliorative encoding method of job based presentation and a corresponding crossover operator are presented. As a result, the developed HGA is fit for the complex problems, that is, it can find an optimal scheduling policy for minimizing the total penalties of early delivery and tardy delivery which occur when jobs are finished outside the due date window. The algorithm is featured by the low decoding complexity, the strong university and robustness. Finally the application of this method is illustrated with examples and presents some problems to be solved on job shop scheduling problems.