A proposal of neuron filter: A constraint resolution scheme of neural networks for combinatorial optimization problems

A constraint resolution scheme in the Hopfield-type neural network named "N euron Filter" is presented for efficiently solving combinatorial optimizati on problems. The neuron filter produces an output that satisfies the constr aints of the problem as best as possible according to both neuron inputs an d outputs. This paper defines the neuron filter and shows its introduction into existing neural networks for N-queens problems and FPGA board-level ro uting problems. The performance is evaluated through simulations wharf the results show that our neuron filter improves the searching: capability of t he neural network with the shorter computation time.