A novel recurrent neural network for solving quadratic programming problems

This paper presents a novel recurrent neural network for solving quadratic programming problems. The new model has a simple form, furthermore has a good convergence rate with a less number calculation operation than the old models. It converges very fast to exact solution of the dual problem and by substituting in a formulation, the optimal solution of the original problem is obtained. We solve neural network model with one of numerical method. Finally, simple numerical example is provided for the sake of illustration.

کلیدواژه ها:

Neural network ، quadratic programming ، differential equation