Apply genetic algorithm to parameter estimation in chaotic noise
Zhenyan Li, Huachun Dong, Taifan Quan · 2003
Minimizing the phase space volume (MPSV) method is a promising method to estimate parameters in chaotic noise, and separate the desired signal from chaotic noise background. However, the high time complexity is a major problem in its algorithm, and this weakness limits its applications. In this paper, we examine the feasibility of using a genetic algorithm (GA) in MPSV, and show the possible decreasing degree of time complexity. To illustrate the usefulness of applying GA, we applied the improved method to estimate the coefficients of an autoregressive model. As we showed, it improves the weakness of the original method, and in our experiment, the time spent by the improved algorithm decreases about 10/sup 2/ times, and maintains the precision at the same time.