A Fractional-order Neural Network for Solving Nonsmooth Programming*

In this paper, a fractional-order neural network is designed to solve non-smooth programming with linear equality constraints. First, we transform the feasible domain of the problem by extreme value mapping. We design a fractional-order neural network model by using the differential inclusion principle. The existence criterion is established by using the KKT condition for the existence of the optimal solution to the original problem. We select the appropriate Lyapunov function to prove the global stability of the equilibrium point of the fractional-order neural network. Finally, a numerical example is given to verify the obtained results in this paper.