A METHOD OF NONLINEAR OPTIMIZATION BASED ON NEURAL NETWORK

This paper describes a specific method of applying neural network approach to solve nonlinear optimization problem. The forming of the energy function is analyzed, and then a neural network dynamic model of nonglobal link on the basis of the regular Hopfield network model is structured. The network improved overcomes the difficulty of the weight map when using regular Hopfield network to solve nonlinear optimization problem. The network structure is distinct,and the software analog and the hardware implementation are easy.