摘要
蚁群算法是一种模仿真实蚂蚁群集体行为的全局启发式随机搜索算法．目前蚁群算法存在易陷入局部最优、搜索时间长等问题。提出一种改进的蚁群算法，加入扰动策略、挥发因子动态调整策略以避免算法陷入局部最优值．采用奖励策略提高搜索效率。通过在旅行商问题上验证得知，改进后的算法可以获得已知最优值，与最大最小蚁群算法相比，解的平均值、出现最优值的概率都有提高。
Ant colony optimization is one of intelligent optimization algorithms from the observations of ant colonies foraging behavior. However, ACO usually costs more searching time and gets into early stagnation during convergence process. Designs an improved ant colony algorithm. It uses perturbation method and adjusts volatilization coefficient to avoid early stagnation and uses hortation method to improve searching efficiency. Applys the algorithm on traveling salesman problem. It shows that the algorithm can find the best value more quickly, has less average value and gets the best value more stability than MMAS algorithm.
出处
《现代计算机》
2011年第12期9-11,共3页
Modern Computer
关键词
蚁群算法
旅行商问题
信息素
扰动
Ant Colony Optimization
TSP
Pheromone
Perturbation
