天车调度是一个NP难的多机多任务调度问题.针对炼钢一精炼一连铸生产过程中的天车调度问题,考虑时间和空间约束,建立了以钢水等待时间最少为目标的天车调度模型.提出了改进的Memetic算法,设计了采用任务分配规则和冲突处理规则的个体解码方式、PPX交叉算子和局域搜索算子;用该算法对模型进行求解,并与遗传算法进行比较.对某钢厂炼钢一精炼一连铸生产过程的天车调度问题的测试表明:与遗传算法相比,用该算法求解的等待时间最多减少16.67%,并具有很高的稳定性.
A crane scheduling problem is a multi-machine, multi-task and NP-hard problem. For the problem in the steelmaking-continuous casting production process, a model was proposed to minimize the waiting time of molten steel, with temporal and spatial constraints considered. An improved Memetic algorithm was developed to solve the model, in which individual decoding based on task allocation rules and conflicts eliminating rules, PPX crossover operator and local search(LS) operator were designed, and it was also compared with the genetic algorithm. The solution to the model with the improved algorithm for a practical problem showed that the waiting time was reduced by 16.67% at most with high stability, in comparison with the genetic algorithm.