摘要
文章在现有动力学演化算法的基础上提出基于混沌的演化算子。除有效地防止过早收敛并且保持解的均匀分布外,新算法充分利用混沌对于初始值敏感性和遍历性的特点,还具有更快的收敛速度和更小的计算量。数值结果显示在解决多峰值问题时新算法不仅有很好的性能和高可靠性,而且优于作者已知的其他已出版的结果。
This paper proposes an improved dynamical evolutionary algorithm （IDEA） based on the theory of statistical mechanics and novel chaotic operation. Besides preventing premature convergence effectively and keeping the population in good distribution, the new algorithm makes full use of initial value sensitivity and track ergodicity of chaos, overcoming the disadvantage of big searching dead zone existed in conventional chaotic mutation model . The numerical results show IDEA not only has good performance and a high degree of reliability while dealing with various complex problems, but also is superior to any other published results known by the authors.
出处
《安阳工学院学报》
2006年第3期52-54,共3页
Journal of Anyang Institute of Technology
关键词
动力学演化算法
过早收敛
混沌初始化
混沌变异
种群多样性
dynamical evolutionary algorithm
premature convergence
chaotic mutation
population diversity
