基于双重阈值检测的自适应人工免疫算法研究

Research on Adaptive Artificial Immune Algorithm based on Double Threshold Detection

人工免疫算法由于抗体记忆库规模固定,在搜索过程中易出现退化、早熟现象。针对该问题,提出了一种基于双重阈值检测的抗体记忆库入库及出库机制,该机制能够保证抗体具有足够的多样性,有效抑制退化、早熟现象以提高算法的全局优化能力。利用常用的算法性能测试函数Two-dimensional Shubert function对所提算法进行性能测验,并与人工免疫算法进行比较。对比分析结果表明,所提算法比人工免疫算法具有更好的全局优化能力和更快的求解速度。

Artificial immune algorithm is prone to degeneration and prematureness in the search process due to the fixed size of antibody memory.Aiming at this problem,a new method of storage mechanism of the antibody memory based on double threshold detection is proposed.This mechanism can ensure that the antibody has sufficient diversity in order to inhibit the degradation and premature phenomenon effectively to improve the global optimization of the algorithm.The commonly used algorithm performance test function-Two-dimensional Shubert function is used on the proposed algorithm for performance testing.Meanwhile,test results are compared with the artificial immune algorithm.The results of the comparison show that the proposed algorithm has better global optimization ability and faster solution than the artificial immune algorithm.