机构地区: 吉林大学计算机科学与技术学院符号计算与知识工程教育部重点实验室
出 处: 《吉林大学学报(信息科学版)》 2003年第4期382-386,共5页
摘 要: 以陆地作战训练模型为背景,研究了多Agent系统中Agent初始属性的优化问题,提出了一种径向基函数(RBF:RadialBasisFunction)神经网络与遗传算法(GA:GeneticAlgorithm)相结合的、对作战训练模型中Agent的初始位置进行优化的方法。与已有的优化方法相比,该方法不仅优化效果得到明显的提高,而且执行效率可以提高20余倍,更适于处理对执行效率要求较高的优化问题。 The initial properties of agents in the MAS(Multi\|Agent System) on the background of land combat simulation model are studied. It proposes a method which combines the RBF(Radial Basis Function) neural network and GA(Genetic Algorithm) to optimize the initial positions of agents in the land combat simulation model. The comparison with the existing method shows that the optimizing results of the proposed method can be increased obviously and the efficiency of the performance can be increased more than twenty times. Therefore, this method is more suitable to handle optimization problems with requirement of speed\|up time.
关 键 词: 多 系统 径向基函数神经网络 支持向量机 遗传算法
领 域: [自动化与计算机技术] [自动化与计算机技术]