基于策略模式的粒子群优化算法平台设计  

Design of algorithm platform for particle swarm optimization based on strategy pattern

作　　者：徐星[1] 李元香[1] 吴昱[1] 胡豪[2] 

机构地区：[1]武汉大学软件工程国家重点实验室,湖北武汉430072 [2]江苏工业学院信息科学与工程学院,江苏常州213016

出　　处：《武汉大学学报（工学版）》2010年第3期361-365,374,共6页Engineering Journal of Wuhan University

基　　金：国家自然科学基金(编号:60773009);国家高技术研究发展计划(863计划)(编号:2007AA01Z290);湖北省自然科学基金(编号:2007ABA009)

摘　　要：为了扩大粒子群优化算法的应用范围和增强它的影响力,从软件重用的角度出发,考虑到算法流程的共性和个性,设计了一种基于策略模式的粒子群优化算法平台,此平台包含了基本的粒子群优化算法和经典的改进算法,可以解决连续优化和二进制组合优化问题.一系列的粒子群优化算法和优化问题被分别封装到相应的算法策略类和问题策略类中,这些类继承自一个具有统一接口的抽象基类.因此,该平台非常适合于粒子群优化算法的理论和应用研究,且易于维护和扩充.In order to extend PSO＇s application area and increase influence of PSO,it is necessary to develop an algorithm platform for particle swarm optimization（PSO）.From the angle of software reuse,we consider universality and individuality of the process of PSOs and design the algorithm platform based on strategy pattern for PSO.The algorithm platform contains basic PSO and many classic improved variants of PSO and solves continuous optimization and binary combination optimization problems.A series of algorithms and optimization problems are encapsulated by algorithm strategy class and problem strategy class respectively.These classes inherit from an abstract base class with a unified interface.So,the platform is very suitable for theoretical and application study;and it is easy to maintain and extend.

关 键 词：策略模式 粒子群优化算法 设计模式 算法平台 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]