基于混沌模拟退火PSO算法的威布尔分布参数估计应用研究

摘要

针对三参数威布尔分布模型采用精确解法直接求解的不足,提出基于混沌模拟退火粒子群优化方法进行参数估计.引入Logistic混沌因子调整粒子群优化算法的更新策略以充分释放其遍历搜索能力,并采用模拟退火方法依据Tsallis接受准则以一定概率接受新状态,使算法避免陷入“早熟”进而实现全局最优搜索;同时为降低算法在迭代计算上的时间开销,运用图解法获得的初始解为其提供搜索范围.将该方法运用到轴承转子可靠度威布尔分布参数估计中,实验分析表明该方法具有可行性和有效性,与遗传算法、模拟退火粒子群优化算法相比具有更好的寻优能力.%Aiming at the deficiency in the precise solution for the three-parameter Weibull distribution model,the parameter estimation based on a chaotic simulated annealing particle swarm optimization algorithm was proposed,and the Logistic chaos factor was introduced to adjust the update strategy of the particle swarm optimization algorithm to fully release its ergodic search ability.The simulated annealing method was used to accept the new state with a certain probability according to the acceptance criteria of Tsallis,so that the algorithm could avoid premature convergence and realize the global optimal search.At the same time,in order to reduce the time of iterative calculations,the initial solution obtained by the graphic method was used to provide the search scope.The method was applied to the reliable Weibull distribution parameter estimation of a bearing rotor.Experimental results show that the method is feasible and effective and has better optimization performance compared with a genetic algorithm and a simulated particle swarm optimization algorithm.