动态分阶段蚁群算法及其收敛性分析

Dynamic stage ant colony algorithm and its convergence

为了提高蚁群算法的收敛速度和求解精度,根据仿生优化算法在不同阶段的特点,提出一种改进的蚁群算法.该算法对参数和选择策略进行了分阶段设计,而且参数的分阶段是根据寻优状态动态划分的.通过对蚁群系统马尔科夫过程进行分析,证明了该算法的全局收敛性.针对典型的TSP问题进行仿真对比实验,验证了该算法在速度和精度方面优于传统蚁群算法.

According to the stage properties of bionic optimization algorithms, an improved ant colony algorithm is proposed to enhance the speed and accuracy of the original algorithm. In the algorithm, parameters and selection Strategy are specially designed in different optimizing stages which are marked referring to the current optimizing states. Its global convergence is proved by analyzing the Markov process of the ant colony system. The contrasting experiments to the typical traveling sales problem prove that the proposed algorithm is advantageous over the traditional ant colony algorithms in speed and accuracy.