Chaotic neural network with nonlinear function self-feedback

Chaotic neural networks can acquire the ability to escape local minima of energy functions by chaotic dynamics. A Novel Chaotic neural network model with nonlinear function self-feedback is proposed by introducing nonlinear function into self-feedback of chaotic neural network. The analyses of the optimization mechanism of the networks suggest that nonlinear function self-feedback affects the original Hopfield energy function in the manner of the sum of the multiplications of nonlinear function to the state, avoiding the network being trapped into the local minima. We constructed the energy function of chaotic neural network, and analyzed the sufficient condition for the networks to reach asymptotical stability and set the parameter set of the networks for solving traveling salesman problem (TSP). Simulation results indicate that the novel chaotic neural networks can find the optimal solution of combinatorial optimization problems.