摘要
提出了一种利用基因表达式编程进行非线性系统辨识的新方法,引入了可变染色体和可变终端符集,提出了新的个体的生成机制及相应进化操作符,克服了利用遗传编程进行非线性系统辨识的不足,降低了算法对参数的依赖性,能够在相同的参数设置下进行各种非线性NARMAX模型辨识。对于适应度的定义,综合考虑模型的精确性和复杂性因素,使辨识模型能够在精确性和复杂性之间取得平衡。仿真结果表明,这种方式可以快速准确的获取非线性模型。
A new method based on gene expression programming for identifying the nonlinear system model was proposed, which introduced variable-length chromosomes and variable-length terminal sets, a new approach was proposed to create initial chromosomes, and genetic operators were revised. It overcame insufficiencies of the initial identifying method based on genetic programming, reduced parameter dependency of evolution algorithm, and could identify various NARMAX models under the same parameters set, The fitness definition considered fully the factors of models＇ accuracy and complicacy, and made the solution get a trade-off between the accuracy and the complexity of the models. The simulation results express that this method can obtain nonlinear model quickly and accurately.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2008年第7期1842-1845,1875,共5页
Journal of System Simulation
