电网节点编号优化的一种改进蚁群算法

An improved ant colony algorithm in electric network node numbering optimization

蚁群算法是近几年优化领域中新出现的一种启发式仿生类并行智能进化系统,目前已经在众多组合优化领域中得到广泛应用.不同于传统的节点编号优化方法,采用最大最小蚁群系统改进的蚁群算法能快速地找到多个全局最优解,并且不易陷入局部最优解.将多种蚁群算法的改进融合在一起,取长补短,得到了较满意的效果.根据节点编号本身的特点,通过引入初始节点选择开关算子,同时在迭代过程中不断淘汰劣势蚂蚁,使蚂蚁能更快地找到最优解.

Ant colony algorithm is a novel category of bionic meta-heuristic system, and has wide application in various combined optimization fields. Different from traditional methods of node numbering optimization, the most global optimal solutions can be found by improved ant colony algorithms based on max-rain ant system, and without dropping in local optimal solutions. Besides, satisfactory result will be obtained if some ant algorithms are used to benefit each other. According to traits of node numbering, a switch operator to choose initial nodes is introduced, and the ants in weaker traces are knocked out in every iterative process, which make the optimal solutions fount out by ants more quickly.