Chaotic Neural Network with Trigonometric Function Self-feedback

A new chaotic neural network is constructed by introducing trigonometric function into the self-feedback term of the transient chaotic neural network; By analyzing the inverse bifurcation graph and the maximum Lyapunov exponent of the single neuron, the chaotic dynamic characteristics of the network are analyzed; The energy function of the chaotic neural network is constructed. Its stability is analyzed and the condition of gradual stability of the network is given; The network parameters for solving the traveling salesman problem are set. The simulation results show that the new model can avoid local minima effectively and is effective for solving combinatorial optimization problems.