利用改进遗传算法的参数估计

基于极大似然法的参数估计实质上是一个复杂的非线性优化问题，传统的优化方法计算效率较低且容易陷入局部极值。而遗传算法是一种有导向的随机搜索方法，能以较大的概率收敛到全局最优解。本文将单纯形法引入到并行遗传算法中，提出了一种改进的遗传算法，可以有效地提高算法的收敛速度、防止搜索过程中的早熟现象。应用于系统初始状态未知时的参数估计问题，获得了满意的结果。

Parameter Estimation Using an Improved Genetic Algorithm

The parameter estimation based on the maximum likelihood method is actually a complicated nonlinear optimization problem The traditional optimization algorithms are apt to be trapped into local minimum and the computation efficiencies are quite low.However,the genetic algorithm is a directed random search method,which has a good probability converge to global optimum.In this paper,an improved parallel genetic algorithm combining with the simplex method is proposed,which could prevent premature convergence effectively and improve the computation efficiency.The proposed algorithm was applied to the parameter estimation problem with unknown initial states of system and satisfactory results are obtained.