一种混合遗传算法及其在线性系统辩识中的应用

提出一种将LMS算法和实数编码遗传算法相结合的混合遗传算法。相对于传统算法，将其应用于线性系统辩识，特别是IIR系统的辩识，具有明显优势。IIR系统的性能函数存在多个局部极小值，该算法首先利用遗传算法的全局最优性，在大范围内搜索可能的极值，并限制系统参数随机搜索的范围，以保证IIR系统的稳定性，再在遗传算法中嵌入一个LMS算子，在极值点附近进行快速搜索，从而达到全局最优与快速搜索的有机结合，提高了辨识精度和计算效率。仿真结果验证了该算法用于线性系统辨识的有效性和实用性。

A Hybrid Genetic Algorithm and Its Application in Linear System Identification

A hybrid genetic algorithm combined with LMS algorithm was put forward.By inserting a LMS operator in Real-coding genetic algorithm (RGA),the algorithm showed good characteristics in IIR system identification. Since IIR system has several local minimums of the objective function,the algorithm first searched for possible minimums of the objective function in a wide range by RGA,limited the random researching range of the system's parameters to ensure stability of the system, and then carried out fast research near the local nadir by LMS operator . The precision of identification and calculating efficiency was improved. Simulation results showed that the algorithm is efficient and useable for the identification of linear system.