基于复数编码遗传算法的竞争性协进化策略

来源期刊:中南大学学报(自然科学版)2005年第3期

论文作者:谭冠政 刘良敏

文章页码:475 - 480

关键词:进化机器人;竞争性协进化;遗传算法;神经网络;连接权

Key words:evolutionary robot; competitive co-evolution; genetic algorithm; neural networks;synaptic weights

摘    要:将基于复数编码的遗传算法引入竞争性协进化的理论研究中,提出一种竞争性协进化的新策略,即:在仿真实验中,采用2个基于神经网络结构控制的移动机器人,并将它们投入到一个陌生的环境中。其中,一个机器人扮演猎手,另一个扮演猎物,猎手对猎物进行捕捉,最终得到每一代的最好猎手机器人和最好猎物机器人以及它们的适应度曲线。在这个竞争性协进化系统中,基于复数编码的遗传算法主要用于对机器人控制系统的神经网络进行进化。计算机仿真结果表明,与基本遗传算法相比,基于复数编码的遗传算法具有更强的进化能力。

Abstract: The genetic algorithm with complex-valued encoding was introduced to the research of competitive co-evolution theory, and a new strategy for competitive co-evolution was proposed. In the simulation experiments, two mobile robots with neural networks control structure were put into an unknown simulation environment. One of them played the part of hunter and the other played the prey; the hunter always tried to catch the prey. The best hunter and the best prey of every generation were obtained from the experiment, including the fitness curves of them. The genetic algorithm with complex-valued encoding was mainly used to evolve the neural networks of controller of robots. The experiment results show that the genetic algorithm with complex-valued encoding has stronger evolutionary capacity compared with the general genetic algorithm.

基金信息:国家自然科学基金资助项目
中国科学院机器人学开放研究实验室基金资助项目

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号