基于离散数字编码的蚁群连续优化算法  

Ant Colony Continuous Optimization Based on Discrete Numerical Encoding

作　　者：吴广潮[1] 黄翰[2] 

机构地区：[1]华南理工大学数学科学学院,广州510640 [2]华南理工大学计算机科学与工程学院,广州510640

出　　处：《计算机科学》2008年第3期146-148,共3页Computer Science

基　　金：国家自然科学基金项目(10471045);广东省自然科学基金(04020079);华南理工大学自然科学基金(B13-E5050190)

摘　　要：本文提出了一种基于离散编码的蚁群连续优化算法(CACO-DE),用于求解连续优化问题。以往蚁群算法(ACO)的研究,以求解离散优化问题为主,较少涉及连续优化问题。与经典的ACO算法不同,CACO-DE将有限精度的实数转化为一个数字串,数字串的每位取0到9之间的数字,从而实现了用离散编码描述实数的效果。CACO-DE延用了经典ACO算法的框架,并加入了特殊的选择机制、信息素更新方式和局部搜索策略。测试实验结果表明:CA-CO-DE比以往同类算法求解速度更快且精度更高。The presented paper proposes an ant colony algorithm for continuous optimization （CACO-DE）. ACO algo-rithms are always used for discrete optimization problems, but rarely for continuous optimization. CACO-DE is designed based on the numerical encoding in which each real number is changed into a string made up of characters { 0,…, 9}. The length of encoding depends on the accuracy and dimension of the solution. Artificial ants construct solutions being guided by a high dimension pheromone vector. The framework of the proposed algorithm is similar to the classical ACO except for the updating rule and local search strategy. Some preliminary results obtained on benchmark problems show that the new method can solve continuous optimization problems faster than other ant and non-ant methods.

关 键 词：蚁群算法 连续优化 离散数字编码 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] O242.23[自动化与计算机技术—计算机科学与技术]