基于混合遗传算法的FMS优化配置及其工艺路线规划

The Configuration and Process Route Optimization of FMS Based on Hybrid GA

针对基于CQN模型的FMS优化配置问题提出了一种混合遗传算法,充分利用CQN模型中生产量函数和成本函数的单调性,设计了最大产量-成本梯度算子来引导新一代种群从不可行域进入可行域.同时,在求解相应的工艺路线规划问题的遗传算法中引入了具有启发式规则的余量随机分配算子,可以将超过约束条件的余量随机分配到个体中去,并通过按照一定规则的调整而保证所有个体的可行性.这样,一方面实现了利用遗传算法求解FMS配置的约束优化问题,另一方面加强了遗传算法的局部搜索能力.算例证明该算法的求解质量好于目前该领域常用的隐枚举算法.

Hybrid genetic algorithms (HGA) was proposed for configuration of FMS based on Closed Queueing Networks (CQN). The monotonicity of throughput function and cost function in the CQN model was fully utilized to design an operator called as Maximum Gradient of Throughput-Cost. This operator can guide new populations into feasible region from infeasible region. The algorithms for process route optimization imbedded an operator called as Operator Distributing Remainder Randomly in GA to resolve the constraint optimal problem. The operator distributed equation constraints over individuals randomly and makes all individuals feasibility by heuristic rules. Accordingly, HGA are capable of solving the fconstraint optimal problem and enhancing their ability of local search. An illustration of the method showed that the solution quality by the HGA is better than that by Implicit Enumeration most in use.