摘要
针对克隆选择算法在求解高维函数优化问题时易陷入局部最优以及收敛速度较慢的弱点,本文基于生物免疫系统内部学习优化机制以及进化算法,提出了一种新的免疫进化算法,它包括正交交叉、单形交叉、克隆、多极变异和选择。新算法将进化计算的思想融入到克隆选择中,提出了一种新的变异算子,在保证种群多样性的同时提高了算法的全局寻优能力。理论分析证明了算法的收敛性,并将算法应用于不同的测试函数进行仿真实验。结果表明,该算法是有效的。
Considering the drawbacks of easily being trapped in a local optimal solution and low convergence velocity of the clone selection algorithms in solving high dimmensional function optimization, this paper proposes a new immune evolutionary algorithm based on the interior learning mechanism of biological immune systems and evolutionary algorithms. The new algorithm includes orthogonal crossover, simplex crossover, clone, multipolar mutation and selection. The idea of evolutionary computation is integrated into clone selection, and a new mutation operator is proposed. This new algorithm can guarantee the diversity of the population and improve the global search ability. Theoretical analyses prove that NIEA converges to the global optimum. Different functions are utilized to test this method and the simulation results suggest that this algorithm has good performance.
出处
《计算机工程与科学》
CSCD
2008年第8期49-52,共4页
Computer Engineering & Science
基金
国家自然科学基金资助项目(60404021
60234030)
国家基础研究项目(A1420060159)
湖南省院士基金资助项目(05IJY3035)
关键词
免疫算法
正交交叉
单形交叉
多极变异
函数优化
immune algorithm
orthogonal crossover
simplex crossover
multipolar mutation
function optimization
