A dual neural network for shortest-path routing
Jun Wang · Proceedings of International Conference on Neural Networks (ICNN'97) · 2002
The shortest path problem is a classical combinatorial optimization problem arising in numerous planning and designing contexts. This paper presents a recurrent neural network for solving the shortest path problem. Based on the dual problem formulation, the recurrent neural network called the dual routing network has much simpler architecture than the one based on the primal problem formulation. The dynamics and architecture of the dual routing network are defined and the operating characteristics are demonstrated via simulation.