基于混合遗传算法的柔性制造系统优化设计  

Optimal Design of FMS Based on Hybrid Genetic Algorithm

作　　者：李建勇[1] 鄂明成[1] 查建中[1] 

机构地区：[1]北方交通大学机械与电子控制工程学院,北京100044

出　　处：《计算机集成制造系统-CIMS》2003年第3期198-201,246,共5页

基　　金：国家自然科学基金资助项目(50175001)

摘　　要：针对基于闭排队网络模型的柔性制造系统优化设计问题,提出了一种混合遗传算法,利用该模型中生产量函数和成本函数的单调性,设计了最大产量一成本梯度算子,来引导新一代种群从不可行域进入可行域,既实现了利用遗传算法求解柔性制造系统约束优化问题,又增强了遗传算法的局部搜索能力。由于该算法利用渐近边界分析思想和编码技术减少了计算量,从而使混合遗传算法既保持了遗传算法的全局寻优特点,又提高了运行效率。算例证明,该算法的求解质量优于目前该领域常用的隐枚举算法。A hybrid genetic algorithm (HGA) is proposed for optimal design of FMS based on Closed -Queuing Network (CQN). The monotony feature of throughput function and cost function in the CQN model is fully utilized to design an operator called as Maximum Throughout - Cost Gradient Operator, which can guide the new populations into feasible region from infeasible region. Thus, the HGA capable of solving the constraint optimal problem of FMS is realized and its ability of local search is enhanced. The method of asymptotic bound analysis and coding technology are integrated into this algorithm to decrease its computing time. Accordingly, the HGA not only inherits GA's global optimization feature, but also is more efficient. An illustration of the method shows that the solution quality by the HGA is better than that by Implicit Enumeration most in use.

关 键 词：柔性制造系统 优化设计 遗传算法 隐枚举法 闭排队网络 

分 类 号：TH165[机械工程—机械制造及自动化]