摘要
人工蚁群算法是一种新型的模拟进化算法,但也存在一些缺点,特别是在规模大的问题中,如计算时间较长、容易陷入局部极小等。本文在基本人工蚁群算法的基础上,引入遗传算子;为保持种群的多样性在选择策略中引入感觉阈值;并根据蚁群蚂蚁是分工的这一思想将其引人蚁群算法中,我们称之为具有分工的蚁群算法,将其应用到TSP问题求解和一些函数优化的实例中。实验结果表明改进是有效的。
Artificial ant algorithm is new in evolving computing, but it has some deficiency, such as its searching speed is slow and it is easy to fall in local peak especially in large scale problem. To overcome these deficiency, we introduced genetic operator and sensation threshold for selection operator to keep the population diversity. We introduced the ants division work to the algorithm we called it the ant colony algorithm with division work and applied it to some examples such as functional optimization and NP-hard problems. The result is obviously improved.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2003年第3期328-333,共6页
Pattern Recognition and Artificial Intelligence
基金
国家自然科学基金(No.60272034
No.60273094)
浙江省自然科学基金青年人才培养专项资金资助项目
关键词
人工蚁群算法
模拟进化算法
函数优化
群体搜索策略
Ant Colony Algorithm, Simulating Evolution Algorithm, Combinatorial Optimization, Function Optimization
