摘要
针对传统蚁群算法存在搜索时间长、易出现早熟和停滞、易陷于局部最优解等缺陷,提出了一种基于凸壳预处理的多态蚁群优化算法。该算法融合快速凸壳查找、多蚁群协同以及信息素扩散等技术,使之更加接近自然界真实的蚁群行为。仿真实验表明,该算法比传统的蚁群算法具有更好的鲁棒性、能跳离局部最优解、收敛速度快、迭代次数少以及全局最优解能力强等特点。
A multi-ant colony optimization algorithm based on convex hull pretreatment is discussed to avoid long- time searching, precocity and stagnation and tendency to local optimization of traditional ant colony algorithm. The multi-ant colony optimization algorithm can make a natural real ant colony by fast convex hull algorithm, polymorphism ant colony coordination and pheromone pervasion algo- rithm. The simulation experimental show that this algorithm has better robustness, fast convergence and less iterative times than traditional ant colony algorithm. And it can easily jump off local optimization and to global optimization.
出处
《计算机工程与设计》
CSCD
北大核心
2009年第11期2775-2778,共4页
Computer Engineering and Design
