Routing applications of the Hopfield neural network
Giovanna Castellano, Giovanni Attolico, Ettore Stella, A. Distante · 2002
a 2-D Hopfield neural network has been used to perform the shortest path task for optimal routing in a packet switched network. The optimum routing problem requires finding the shortest path in a graph as a sub-task. This optimization problem can be solved using the Hopfield neural model (an interesting alternative to classical optimization techniques), strongly sensitive to the tuning of coefficients in the energy function which maps the problem constraints. The paper presents reliable guidelines, derived from an extended experimental analysis, for fixing these coefficients and simulation results of the neural routing algorithm.