Guided Maximum Entropy Method Algorithm for the Network Topology and Routing

This paper presents an algorithm that applies a guided maximum entropy method to the network design problem. Network design problem is a well known NP-hard problem which almost always involves underdetermined systems, especially when routing policy has to be determined. The maximum entropy method is a relatively new technique for solving underdetermined systems. We adjusted the network design problem, primarily the routing feasibility, to the maximum entropy method requirements. Computationally feasible algorithm is developed which includes additional constraints that direct uniformity of the solution in the desirable direction. Proposed algorithm computes a reasonable solution that is robust with respect to often required dynamic changes of the cost function. This modified method exploits the property of the MEM that it can smoothly move from cases where constraints can be satisfied to cases where constraints become desirable goals that are satisfied as much as possible. A software system was developed which includes all the mentioned features.