Topological design of interconnected LANs using Hopfield neural networks

This paper describes the use of Hopfield neural networks for designing LAN-LAN interconnection with the objective of minimizing the average network delay. More specifically, if we are given a number of transparent bridges to interconnect LANs into a spanning tree structure, we address the issue of placing the bridges so as to minimize the average network delay. The above problem is a complicated combinatorial problem. We use the Hopfield neural network technique to find a heuristic solution. This neural network technique can also be applied to the physical topological design problem in a given LAN-LAN interconnected topology as well