改进克隆选择算法的层叠滤波器的优化设计

PDF下载HTML

Title:

Optimal design of stack filters using an improved clonal selection algorithm

文章编号:

1673-4785（2008）03-0254-05

作者:

赵春晖; 赵华; 万建;

哈尔滨工程大学信息与通信工程学院; 黑龙江哈尔滨;

Author(s):

ZHAO Chun-hui; ZHAO Hua; WAN Jian

College of Information and Communication Engineering; Harbin Engineering University; Harbin 150001; China

关键词:

层叠滤波器;  多克隆算子;  克隆选择;  变异;  重组

Keywords:

stack filters;  polyclonal operator;  clonal selection;  mutation;  regrouping

分类号:

TN911.26

文献标志码:

A

摘要:

层叠滤波器优化设计的核心是正布尔函数最优化的问题.为了提高优化速度并对层叠滤波器进行全局优化设计,提出了一种改进的克隆选择算法.该算法引入多克隆算子和记忆单元及保留群体同时进化的思想.多克隆算子中的个体克隆规模根据个体的亲和度浓度自适应变化;重组操作在父代记忆个体与子代变异后的记忆个体之间展开,避免了近亲繁殖;保留群体的变异保证了群体的多样性.实验结果证明,该算法优化的层叠滤波器能在较短的时间内得到较好的滤波结果.

Abstract:

The op timal design of stack filters is actually an op timization of a positive Boolean function. In order to speed up the op timizing rate and p roduce a global op timal design of the stack filters, an imp roved clonal selection algorithm ( ICSA) is p resented, which introduces a polyclonal operator and simultaneous evolution of the memory u2 nit and the reserved group. The number of clonal particles in the memory unit can adap tively change with their af2 finity concentration. Regroup ing operations are made between parent generations and child generations in the mem2 ory unit, avoiding p roblems caused by breeding two close particles. The p reservation of group mutations guarantees diversity. Our experimental results confirmed that stack filters op timized with ICSA p roduce better filtering results in less time.