An Adaptive Search Space Based Evolutionary Algorithm with Application to Actuator Hysteresis Identification

摘要

This paper presents the approach of using an Evolutionary Algorithm with Adaptive Search Space (EAASS) in identifying the hysteresis parameters of an electromechanical valve actuator. The proposed EAASS features an adaptive mechanism to control the search space as well as the rate of crossover. According to the normalized fitness distance in each generation, EAASS consistently identifies the best search domains in the parameter space and adjusts the crossover rate in order to improve the solution accuracy. To further enhance the robustness, EAASS allows the crossover rate to grow exponentially and the mutation rate to decay logarithmically according to the generation number. The hysteresis model of a simulated valve actuator identified by EAASS has shown high accuracy.