Simulated Annealing Strategy in Chaotic Neural Network with Chebyshev Polynomials

By combining the Chebyshev polynomials and the Sigmoid function into a new non-monotonic excitation function, a new transient chaotic neural network model is constructed. The dynamic characteristics of the single neuron are analyzed by the time evolution graph and the inverted bifurcation graph of the largest Lyapunov exponent. Verify the rationality of the additional energy term of the constructed model. The effectiveness of the model is verified by nonlinear function optimization and traveling salesman problem. The simulation results show that the newly constructed model can effectively avoid the problem of converging to a local minimum in the optimization process, and can effectively solve the combinatorial optimization problem.