基于混沌混合算法的装配序列规划

发布时间：2018-11-20 15:29

【摘要】：针对遗传算法在求解装配序列规划问题中速度慢、产生重复解等问题,提出一种将蚂蚁算法、混沌算法和遗传算法结合,加入动态更新种群数目这一策略的新方法。利用拆卸干涉矩阵获取零件之间的信息,建立评优指标的信息矩阵。算法前期采用蚂蚁算法进行局部搜索,后期运用混沌—遗传算法进行全局搜索。在全局搜索中,采用Logistic映射生成子代装配序列以避免产生重复解,并设置一定的比例筛选搜索结果,动态更新种群数目。为进一步提高序列优劣的区分度,将装配工具的运用次数与装配方向的改变次数合并为连贯性,同时引入辅助工具这一指标。以齿轮油泵为例,将所提算法与蚂蚁算法、遗传算法和混合算法进行比较,从适应度值、运行时间和收敛速度方面进行分析,验证了该算法的有效性。
[Abstract]:Aiming at the problem of slow speed and repeated solution in solving assembly sequence planning problem by genetic algorithm, a new method which combines ant algorithm, chaos algorithm and genetic algorithm to dynamically update population number is proposed. The disassembly interference matrix is used to obtain the information between parts, and the information matrix of the evaluation index is established. Ant algorithm is used for local search and chaos-genetic algorithm is used for global search. In the global search, the assembly sequence of offspring is generated by Logistic mapping to avoid repeated solutions, and a certain proportion is set to filter the search results and dynamically update the population number. In order to improve the classification of the sequence, the combination of the number of the assembly tools and the changing times of the assembly direction into the coherence, and the introduction of the auxiliary tool as the index at the same time. Taking the gear pump as an example, the proposed algorithm is compared with ant algorithm, genetic algorithm and hybrid algorithm. The validity of the proposed algorithm is verified in terms of fitness, running time and convergence speed.
【作者单位】：桂林电子科技大学广西可信软件重点实验室;
【基金】：国家自然科学基金资助项目(61100025,61262030,61363030) 广西自然科学基金资助项目(2014GXNSFAA118354)~~
【分类号】：TG95;TP18

【相似文献】