基于自适应遗传算法的Weibull分布参数评定方法

Parameter Evaluation of 3-parameter Weibull Distribution based on Adaptive Genetic Algorithm

鉴于以三参数Weibull分布建立可靠性寿命模型中存在参数评定误差太大、效率不高的问题,首先建立基于极大似然函数估计的极似然方程组,其次针对传统的遗传算法求解极大似然方程组过程中各个步骤,提出相应的优化与改进,然后得到自适应遗传算法来更精确地求解极大似然函数方程组。最后,通过MATLAB仿真对比分析,将自适应遗传算法与传统遗传算法求解结果进行对比,得出自适应遗传算法在求解Weibull分布参数中具有更高的效率以及适应性。该方法对基于极大似然函数估计的可靠性寿命模型的求解提供了参考。

The study is aimed at solving the problems of error and inefficiency in parameter evaluation of a reliability life model in 3-parameter Weibull distribution.Firstly,the maximum likelihood equation was made in the study.Then the deficiency,in steps of using traditional genetic algorithm to solve the maximum likelihood equations,was optimized.An adaptive genetic algorithm was obtained that can be used to solve the maximum likelihood equations.Finally,it compared the simulation results in MATLAB between adaptive genetic algorithm and traditional genetic algorithm.It can be concluded that the adaptive genetic algorithm was more efficient and adaptable than the traditional genetic algorithm.This method also provides a reference for solving the similar problems in parameter evaluation.