解排列优化的整数编码多智能体进化算法.PDF

软件2011年第32卷 第5期 Software 国际IT传媒品牌 解排列优化的整数编码多智能体进化算法 袁 志 (广州大学华软软件学院，广州 510990) 摘　要： 为解排列优化问题，在多智能体进化算法的基础上，提出一种整数编码的多智能体进化算法。重新定义了竞争算子和 自学习算子。在网格内，智能体与周围的8个智能体构成竞争域，优胜智能体将编码段植入失败智能体，只有优胜者能获得自学习 机会，自学习算子中智能体通过两种编码段换位方式来提升能量。使用本算法在旅行商问题典型数据上进行测试，与现有文献比较， 表明该算法具有更好的全局寻优能力而且收敛稳定性更好。 关键词： 排列优化；多智能体进化算法；旅行商问题；进化算法；整数规划 中图分类号： 文献标识码： TP18　　　　　　 A　　　　　　DOI: 10.3969/j.issn.1003-6970.2011.05.007 Number Coding Multi-Agent Evolutionary Algorithm for Deployment Optimization YUAN Zhi (South-China Institute of Software Engineering, Guangzhou University, Guangzhou 510990, China) Abstract For deployment optimization, Number Coding MAEA named as NC-MAEA is proposed based on Multi-agent 【 】 Evolutionary Algorithm, the competition operator and learning operator are redefined. In grids, a agent and 8 agents around constitute a competing area, the winner plants it’s codes fragment into failure’s codes series, only winner obtains learning chance, in learning process it swap code fragments in two ways to improve energy. Simulated with the typical test data of in traveling salesman problems, compared with other related literatures, it has been proved that NC-MAEA has better global optimization ability and convergence stability. Key words deployment optimization; Multi-Agent Evolutionary Algorithm; traveling salesman problem; Evolution Algorithm; 【 】 integer programming 0　引　言 法，证明了算法的全局收敛性。但以该算法求解排列优化问题 时存在明显不足：采用 0-1 编码产生大量重复编码段属于非法 排列优化问题广泛应用在生产调度、工程规划等方面。不 解，导致大量无效计算；0-1 编码使解空间编码规模庞大，迭代 失一般性，