【作者】 张伟伟；

【导师】 周永权；

【作者基本信息】 广西民族大学 ， 计算数学， 2010， 硕士

【摘要】 人口迁移算法是近年来提出的一种新型群智能方法,其理论和算法还不够完善,需要人们不断提出新的理论和方法对其改进。基于此,本文将从人口迁移算法理论框架描述和应用方面进行研究,首先依据群智能优化算法搜索机制的特点构建出人口迁移算法的基本理论框架,对理论框架进行数学化描述,这样可以有助于系统全面的理解算法的原理和本质,有助于今后进一步研究人口迁移算法的相关理论;接着引入混沌优化搜索技术和偏序关系对其进行改进,前一种主要是在算法框架下加入混沌机制改进算法,并通过一算例来验证算法的有效性和正确性,后一种是改进算法的评价方式,用偏序关系取代适应度函数,并举例分析算法的可行性,另外偏序关系为算法的收敛性提供了一定的保证;最后把人口迁移算法应用到径向基函数神经网络的权值训练中,从实验结果分析可看出基于人口迁移算法训练权值方法比RBF算法具有较快的收敛速度和较高的计算精度,验证人口迁移算法的全局优化能力。更多还原

【Abstract】 Population Migration Algorithm is a new swarm intelligent algorithm proposed recently. Though many numerical experiments illustrated its effectiveness and correctness, its theories and algorithms is still not consummate. People need propose new theories and methods to improve it. As research content, this paper will describe the theoretical framework of population migration algorithm and its application. First this paper describes the PMA based on a given framework of swarm intelligence, it helps to understand the principles of the PMA, as well as combination with chaos theory, a Chaos Population Migration Algorithm is proposed, a numerical experiment example indicates that the new algorithm is effective and accuracy, then with the partial order to replace the fitness function ,and use a example to analysis the algorithm ,else the partial order can make sure of the convergence of the algorithm; Last, this paper try to use algorithm for the weight optimization of RBF neural networks and a model based on this method were presented here, compared with the traditional RBF algorithm, the simulation study indicates that this new method is of high convergence precision and convergence precision.更多还原