Chaotic neural networks with Gauss wavelet self-feedback and their applications to optimization

This paper proposes chaotic neural networks with nonlinear Gauss wavelet self-feedback. Chaotic neural networks with wavelet self-feedback not only have the ability of globally searching optimum due to chaos but also have the ability of local approximation due to wavelet. The analyses of asymptotical stability demonstrate the proposed networks can converge stably. The experimental results show that the performance of chaotic neural networks with Gauss wavelet self-feedback is superior to those only with linear self-feedback.