Research On Improved Intelligent Optimization Algorithm Based On IIR Filter Identificatio

IIR filters,infinite impulse response filters,are an important class of filters.In many application fields such as geological prospecting,voice and image signal processing,and radar signals,IIR filters play a very important role and show their huge application value.Traditional IIR filter design methods have serious distortion or spectrum aliasing.By applying intelligent optimization algorithms to identify unknown IIR filter systems,designing corresponding IIR filters has become a research hotspot.Since the intelligent optimization algorithm can effectively solve the identification problem of the IIR filter system,it is necessary to design an intelligent optimization algorithm with better identification effect.This paper proposes two improved rat swarm optimization algorithms: multiple adaptive strategies-based rat swarm optimizer(MARSO)and the opposition-based learning DE-RSO hybrid optimizer(OBLDE-RSO).MARSO uses adaptive learning exemplars strategy(ALE)and adaptive population size strategy(APS),these two strategies can improve the accuracy of the algorithm and save computing resources;OBLDE-RSO uses DE-RSO hybrid strategy and OppositionBased Learning strategy(OBL),these two strategies can expand the search range of the algorithm to find the optimal solution faster.These two improved algorithms and other algorithms are compared with the IEEE CEC2017 benchmark function set.The tests of classic intelligent optimization algorithms and the comparison of their results prove that the two improved algorithms have better performance than other algorithms.At the same time,these two system identification schemes of improved rat swarm optimization algorithm are applied to the selected three benchmark IIR filter models,and compared with the experimental results of other optimization algorithms in solving this problem,through multiple indicators The above comparison proves that the two improved algorithms have excellent performance in the recognition of several IIR filter models.