摘要
本文在蚁群系统的基础上,提出一种改进型蚁群算法。蚂蚁之间通过外激素进行间接交流从而达到合作的目的,在利用已有信息与探索新解并重的策略指导下给出所求解问题的最优解,并且由于遗传算子的引入及全局更新规则的修正,不再易于陷入局部极小。本文采用改进型蚁群算法求解复杂的组合优化问题一旅行Agent问题,取得了满意的效果。实验结果表明,改进型蚁群算法具有鲁棒性强、自适应、并行化、正反馈的优点。
An improved ant colony algorithm on the basis of ant colony system is presented in this paper. Ants cooperate using an indirect form of communication mediated by pheromone and find good solutions to their task guided by the tradeoff between exploitation and exploration. The probability of premature convergence is low due to the introduction of genetic operator and modification of global updating rule. The traveling agent problem is defined in this paper. It is solved using improved ant colony algorithm and promising result is obtained. Experimental results show that improved ant colony algorithm exhibits some excellent characteristics of robustness, self-adaptation, parallelism and positive feedback.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2003年第1期6-11,共6页
Pattern Recognition and Artificial Intelligence
基金
教育部重点项目(No.00071)
安徽省自然科学基金资助项目(No.00043235)