基于改进差分进化算法的IIR滤波器设计

摘要

The Differential Evolution Algorithm is concerned for its properties of implementing simplified, convergence fast and requiring less features. Generally, according to constraint conditions of the optimization problems in differential evolution algorithm, the boundary of arguments are required to be restrained. In this paper, an improved algorithm which using the features of zero and pole points of IIR filters is developed for the design of IIR filters. A complex coding method is used to decrease the dimension of arguments. The solutions out of constraint conditions are used to implement mutation operation to achieve the goal of faster and more accurately converge to the global optimal solution.%差分进化由于其算法简单、收敛速度快，所需领域知识少而受到关注。通常，根据优化问题的约束条件差分进化算法需要进行变量上、下界的限制。本文提出利用IIR滤波器零极点特性来进行IIR滤波器设计，采用复数编码降低了变量的维度，并在设计IIR滤波器时不需要对变量范围进行约束，同时利用超出边界的解进行变异操作，以达到更加快速精确的收敛到全局最优解的目的。