【摘要】 针对混合蚁群遗传算法容易融合时机过早或过晚、种群进化经历的代数过多、效率低等问题,首先改进了蚁群算法,并将改进的蚁群算法和遗传算法结合,应用于Agent联盟求解。提出了基于混合蚁群遗传算法的Agent联盟求解算法(Hybrid Ant Colony and Genetic Algorithm,HAGA),算法的核心是动态寻找两个算法的衔接点,在该点左侧使用遗传算法,右侧使用蚁群算法。与其他传统算法的实验比较,证明了该算法在求解联盟的最优解的时间和精度上都有较高的效果。把HAGA应用于RoboCup2D龙队客户端程序中,使用比赛分析工具软件SoccerDoctor对比赛结果进行了统计分析,结果显示龙队在诸多技术参数方面均占有明显优势。更多还原

【Abstract】 Aiming at such problems as too early or too late fusion of the hybrid ant colony,too many generations of the species evolution and low efficiency,the ant colony algorithm,first of all,got improved.And the connection of the improved ant colony algorithm and the genetic algorithm is applied to the problem-solving of the Agent-coalition.The algorithm of the agent coalition was put forward.The core to the algorithm dynamically searches for the joint point of these two algorithms in a dynamic way.The genetic algorithm is used in the left of this point,the ant colony algorithm in the right.Compared with the traditional algorithm experiments,these algorithms is highly efficient in the time and precision of the algorithm of the coalition.The HAGA was used in client end of RoboCup 2D Dragon Team in Jiang Su University.The match result was analyzed statistically by the analysis tool software—SoccerDoctor in China University of Science and Technology.The result shows that the Dragon Team has evident advantage in many technology parameters.更多还原