考虑热处理批量的模具制造网格服务配置算法

Mould manufacturing grid service allocation algorithm considering heat batch problem

在制造网格环境中,从项目的多任务和项目网状约束的宏观角度出发,考虑在制品物流的情况、加入热处理批量调度问题以及由此问题所引起的项目网络结构的改变情况,研究了模具项目的服务资源优化配置问题。结合遗传算法与模拟退火算法提出一种两阶段的服务优化配置方法。第一阶段通过服务的搜索与匹配得到任务的候选服务节点集合,第二阶段采用混合遗传算法从项目的宏观角度对候选服务节点进行统一优化与配置。使用模拟退火处理热处理任务的批量划分与服务配置问题,使用遗传算法处理非热处理任务的服务优化配置问题。设计了染色体的编码方式,选择、交叉、变异算子,以及热处理任务的模拟退火批量解决方法。通过实例验证了该算法的有效性。

In the manufacturing grid environment, proceed from macroscopic perspective of item＇s multi-task and reticular constraint, the mould project service resource optimal allocation was studied by considering Work In Process （WlP） material flow in manufacturing grid, the heat treatment batches scheduling problem and the project grid＇s structure change. A two-stage service optimal allocation method combined genetic algorithm with simulated annealing algorithm was proposed. In first stage, candidate service node sets through service searching and matching were obtained; In second stage, the candidate service nodes were optimized and allocated unifiedly from macroscopic perspective by using hybrid genetic algorithm. Simulated annealing algorithm was used to handle the batch division and service allocation of heat treatment tasks, and the genetic algorithm was used to deal with the service optimal allocation of non-heat treatment tasks. The solutions of chromosome coding method, selecting, crossing, and mutating as well as the method of batch treatment were designed. The example was used to verify the effectivenesss of proposed algorithm.