论文部分内容阅读
基于克隆选择原理,引入混沌机制和小生境技术,提出一种改进型克隆选择算法(ICSA).该算法比传统的克隆选择算法具有更好的种群多样性和全局寻优能力.以随机过程理论为数学工具,分析了ICSA所形成抗体种群的平均适应度函数的鞅性质,并由此得出算法几乎处处强收敛性的结论.进而证明了,当状态空间有限时,该算法能在有限步内以概率1收敛到全局最优.仿真实验表明,该算法能有效地抑制早熟,具有更好的全局收敛性.
Based on the theory of clonal selection, chaos mechanism and niche technology are introduced, an improved clone selection algorithm (ICSA) is proposed, which has better population diversity and global optimization ability than traditional clonal selection algorithm.According to stochastic process theory As a mathematical tool, the martingale properties of the average fitness function of the antibody population formed by ICSA are analyzed, and the conclusion that the algorithm almost everywhere is strong convergence is obtained. Then it is proved that this algorithm can be used in finite-step Which converges to the global optimum with probability 1. Simulation results show that this algorithm can effectively restrain prematurity and has better global convergence.