摘要
目前,对于整体优化问题已经进行了大量理论研究,并提出了许多基于导数的解析方法和其他非解析的数值优化技术。但是,在实际领域中存在着各种高度复杂的优化问题,其目标函数可能表现为非连续或非处处可微、非凸、多峰和带噪声等各种形式,这类复杂优化问题不适合于采用解析方法,同时用传统上的搜索技术求解也会遇到许多困难。针对上述问题,提出利用遗传算法求解多峰函数的优化方法,新方法利用遗传算法的鲁棒性,对多峰函数进行优化,并用Matlab进行仿真,实验结果表明,遗传算法可以快速稳定地搜索到多峰函数的最优解。
At present,the overall optimization problems for a large number of theoretical studies have been carried out,and many analytical methods based on derivative and other non-analytical numerical optimization techniques have been presented.However,there are various of highly complex optimization problems existed in actual field,the objective function may be non-continuous,non-differentiable everywhere,non-convex,multimodal and with various forms of noise,the analytical methods are not suitable for such complex optimization problems and it is difficult for using the traditional search technology.In response to these problems,the Genetic Algorithm is proposed for solving Multimodal Function Optimization.The new method uses the robustness of genetic algorithm,optimizes multi-modal function,and performs simulation with Matlab.The results show that the genetic algorithm can search for a stable multimodal function optimization.
出处
《现代电子技术》
2011年第16期68-69,74,共3页
Modern Electronics Technique
基金
国家自然科学基金资助项目(60963016)
关键词
遗传算法
多峰函数
优化方法
鲁棒性
genetic algorithm
multimodal function
optimization
robustness
