Abstract: Aiming at the unidirectional guided-path network design problem of automated guided vehicle system, a Hybrid Genetic Algorithm(HGA)was proposed. The formal model considering both empty and loaded travel distance was given, and the guide-path network was preprocessed to reduce computation complexity. HGA used binary coding, and each binary chromosome determined the direction of a unidirectional segment. To improve the convergence speed of Genetic Algorithm(GA), tabu search was adopted after selection, crossover and mutation operation. To keep a high degree of population diversity, Hamming distance based Niche competition operation was adopted when a new population was generated. The experimental results showed that the proposed HGA was more efficient than traditional GA and tabu search algorithm.

Key words: automated guided vehicle system, hybrid genetic algorithms, tabu search, unidirectional guided-path network

摘要： 针对自动导引车系统单向导引路径网络设计问题提出一种混合遗传算法。建立综合考虑空载和负载总路程的路径网络设计模型，并对路径网络进行预处理,以降低算法复杂度；遗传算法采用二进制编码，每位染色体都对应一条路径的方向；为提高遗传算法的收敛速度，在选择、交叉和变异后增加了禁忌搜索操作；为保持种群的多样性，在形成新一代种群时采用基于海明距离的小生境淘汰运算。实验结果表明，与传统遗传算法和禁忌搜索算法相比，所提算法具有更好的整体性能。

关键词: 自动导引车系统, 混合遗传算法, 禁忌搜索, 单向导引路径网络