维修时点预测的动态车间调度问题

针对制造业中生产计划的不确定问题，提出一种维修时点预测与自适应的遗传模拟退火算法相结合的优化调度方法。该方法首先利用差分自回归移动平均模型预测设备未来的故障率，然后借助电气设备的威布尔（Weibull）分布模型逆向求出设备未来故障发生时刻，最后将此作为约束条件，利用自适应的遗传模拟退火算法解决传统的生产调度问题。结合工厂实际情况，主要分析了设备有无维修的随机调度问题，以最小化最大完工时间为目标，获取每一个任务的调度计划以及每一台设备的维修时点，确定出最佳调度方案。实验表明自适应的遗传模拟退火算法的性能较好。在河北某工厂的生产车间中，设备在运行调度方法后三个月的平均故障率比运行前相对降低了3.46%。

Dynamic shop scheduling problem of maintenance point prediction

Aiming at the uncertain issue of the production plan in manufacturing industry, an optimal scheduling method which combined the prediction of maintenance point with the adaptive algorithm of genetic and simulated annealing was proposed. First of all, the Auto Regressive Integrated Moving Average model (ARIMA) was used to predict equipment failure rate; then the Weibull distributed model was used to reverse the equipment maintenance point in the future by equipment failure rate; finally, regarding the maintenance point as a constraint condition, the traditional production scheduling problem was solved by the adaptive hybrid algorithm of genetic and simulated annealing. The random scheduling situation of equipment for maintenance was analyzed in combination with the practical situation of the factory, and to determine the optimal scheduling scheme, the minimum makespan was regarded as a goal to obtain the scheduling plan of each task and maintenance point of each equipment. Experimental results show that the adaptive genetic and simulated annealing algorithm has good performance. In the production workshop of a certain factory in Hebei, the average failure rate of the equipment which used optimization scheduling method was relatively reduced by 3.46 percent than that before optimization.