An Improved Internal Dynamics in Binary Neural Network for Solving Optimization Problems

In this paper, an improved internal dynamics in binary neural network is proposed for efficiently solving combinatorial optimization problems. The improved internal dynamics provides a mechanism for increasing the exchange of information between neurons, by using a stabilization parameter to control the updating of neurons inputs. The performance was evaluated through simulating the N-queens problem. Simulation results showed that the proposed technique has improved the searching capability of the network and decreased computation time