一种免疫算法在故障诊断中的研究

Research of an Immune Algorithm in Failure Detection

在原有遗传算法的交叉,变异的基础上,将基本免疫算法与遗传算法相结合,通过“精英策略“和“改善变异因子”提高了算法的性能,并将改进后的算法运用到免疫算法中,提高了抗体的适应度和生成机制,其次引入支持向量机,将改进后的免疫算法寻求最优的支持向量机参数,提高了故障的识别率。仿真实验以无线传感器故障为例,在均绝对误差,均方误差和均方根误差等方面都优于基本免疫算法。

Firstly,the basic immune algorithm is combined with genetic algorithm based on the original crossover and mutation of genetic algorithm,and "elite strategy"and "improved mutation factor"have improved the algorithm’s performance,which is used in the immune algorithm to improve the antibody’s adaptation and generation mechanism. Secondly, supporting vector machine is introduced and the improved immune algorithm is used to seek the optimal supporting vector machine ’s parameters,improving the failure identification rate. The simulation experiment takes wireless sensor failure as an example and is superior to basic immune algorithm in mean absolute error,mean square error and root mean square error.