连续区域改进蚁群算法的研究

Improved Research on Ant Colony Optimization in Continuous Spaces

为了解决蚁群算法难处理连续区域的问题,本文结合微粒群操作改进蚁群算法。采用平均分割定义域的方法,融入随机操作和微粒群操作的交叉应用,并加入了信息素的变异操作跳出停滞状态。该混合群算法同时具有全局寻优特性和较强的局部搜索能力,在确保全局收敛性的基础上,能够快速搜索到高质量的优化解。通过仿真算例分析了其可行性、优越性。

As the ant colony algorithm can not use in continuous spaces, an improved ant colony algorithm by particle swarm optimization is proposed. This algorithm divides the total domain averagely, and integrates random operations with particle swarm optimization, and uses the variation operation of pheromone to jump algorithm stagnation. The mix-optimization owns the large-scale search and the local search ability, it also can search the high quantity optimization solution and ensure global convergence. And the result about the emulated test demonstrates the mix-optimization^s possibility and superiority.