An enhanced Benders decomposition method for unique shortest path routing

Abstract

With the aim of looking into a comparatively more efficient solution approach for the unique shortest path routing problem, an alternative problem formulation to a counterpart is first introduced. The model is mathematically equivalent to the counterpart concerning both the feasibility and optimality of the problem. By exploring the constraint structures and model sizes of the two formulations, the former is identified to be a more appropriate one for a decomposition algorithm to solve the problem. An enhanced Benders decomposition method, embedded with redundancy generation strategy, is then proposed, based on the study of problem properties of network flow problems.