优化算法的信息原理与群搜索  

Information principle and swarm search of optimization algorithm

作　　者：莫愿斌[1] 刘贺同[1] 

机构地区：[1]广西民族大学数学与计算机科学学院,广西南宁530006

出　　处：《计算机工程与设计》2008年第4期909-913,共5页Computer Engineering and Design

摘　　要：从两个不同的角度对各种进化算法进行剖析,寻求合理的解释。从信息论的角度进行讨论,分析了在算法设计中如何表示信息、提取信息、传递信息、利用信息与信息融合等问题,在优化技术与信息论之间建立了联系的纽带。从搜索的角度进行讨论,分析了单点搜索与多点搜索,指出了各种进化算法均属于随机布点,企图寻找全局最优的数值方法。Aimed at various evolutionary algorithm being research hotspot now, two different points of view are given to dissect them, in order to find reasonable interpretation for them. Firstly, based on information theory, the information expression, abstraction, transmittal, utilization and fusion in the algorithm design are analyzed, so it found a relation between optimization algorithm and infor- mation theory. Secondly, various evolutionary algorithms fi＇om search are analyzed, and it is drawn that each evolutionary algorithm is a numerical algorithm that randomize initial points and try to find goal optimization.

关 键 词：信息 信息方法 信息融合 优化算法 单点搜索 群搜索 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]