A Neural Network Approach to Routing in Multihop Radio Networks

Although the issues of routing and scheduling in packet radio networks are highly interdependent, few studies have addressed their interactions. In this paper, we address the problem of routing for the minimization of congestion as a first step toward the solution of the joint routing-scheduling problem. We formulate this as a combinatorial-optimization problem, and we develop a Hopfield neural network (NN) model for its solution. The determination of the coefficients in the connection weights is the most critical issue in the design and simulation of Hopfield NN models. In our studies, we use the method of Lagrange multipliers, which permits these coefficients to vary dynamically along with the evolution of the system state. Extensive software simulation results demonstrate the ability of our approach to determine good sets of routes in large, heavily-congested networks.