摘要
针对蚁群优化(ACO)只适用于离散问题的局限性,提出了连续蚁群优化算法(CACO),保留了连续问题可行解的原有形式,并融入演化算法(EA)的种群与操作功能.CACO将蚁群分工为全局和局部蚂蚁,分别引领个体执行全局探索式搜优与局部挖掘式搜优,并释放信息素,由个体承载,实现信息共享,形成相互激励的正反馈机制,加速搜优进程.实例测试表明,CACO适用于连续问题,全局寻优性能良好,尤其对复杂的高维问题,更能反映其相对优势.最后讨论了局部寻优方法、全局蚂蚁配比、挥发因子和种群规模等因素对CACO寻优性能的影响.
Aiming at the shortcoming of ant colony optimization （ACO） which can only apply to discrete problems, a new ACO algorithm for continuous problems （CACO） was put forward. CACO combines ACO with evolutionary algorithms （EA）. It merges both population and genetic operations concepts of EA. It preserves the original form of feasible solutions of the continuous problems. In the proposed algorithm, the ant colony is divided into global ants and local ants, which do the global exploratory optimization and local exploitation optimization respectively. Ants deposit pheromone on the individuals which they selected and share the information in the ant colony, which leads to the positive feedback mechanism and accelerates the optimization process. Experimental results show that CACO is fit for continuous optimization, and that it has advantage over standard genetic algorithm （SGA） in global optimization, especially when applied to high dimensional complex problems. The main influencing factors such as the optimization methods which local ants use, global ants ratio, evaporation factor and size of regions were also discussed.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2005年第8期1147-1151,共5页
Journal of Zhejiang University：Engineering Science
基金
国家自然科学基金资助项目(20276063).
关键词
蚁群优化
演化算法
信息素
探索性
挖掘性
全局寻优
ant colony optimization
evolutionary algorithm
pheromone
exploration
exploitation
global optimization
