论文部分内容阅读
本文提出一种利用遗传算法的最优正交小波基选择方法。通过将信号用小波级数展开后得到其在某个期望尺度上的近似表示，由此建立一个表达信号与其近似之间误差的代价函数，然后我们利用遗传算法最优化此代价函数以获得全局最优的正交小波基。实验结果表明本文的方法能够快速准确地得到全局最优的正交小波基，从而得到信号在期望尺度上的最佳近似。
This paper presents a selection of optimal orthogonal wavelet bases using genetic algorithms. By developing the signal by a wavelet series, we obtain an approximate representation of the signal on a desired scale, thereby establishing a cost function that represents the error between the signal and its approximation. We then use genetic algorithms to optimize this cost function to obtain the global optimum Excellent orthogonal wavelet base. The experimental results show that the proposed method can get the globally optimal orthogonal wavelet bases quickly and accurately, and obtain the best approximation of the signal on the desired scale.