田中 衞

The maximum-flow neural network(MF-NN) is a novel neural network model for the maximum flow problem. From the max-flow min-cut theorem, it is known that the maximum flow problem and the minimum cut problem are dual problems. This indicates that MF-NN is applicable to the minimum cut algorithm. In this paper, we propose a novel minimum cut solution using MF-NN in directed and undirected graphs. Furthermore, since the proposed method is intended to circuit implementation based on nonlinear circuit theory, it has considerable potential for speeding up computation time.