A Neural Benders Decomposition for the Hub Location Routing Problem

Rahimeh Neamatian Monemi,Shahin Gelareh

2023-09-03

Abstract:In this study, we propose an imitation learning framework designed to enhance the Benders decomposition method. Our primary focus is addressing degeneracy in subproblems with multiple dual optima, among which Magnanti-Wong technique identifies the non-dominant solution. We develop two policies. In the first policy, we replicate the Magnanti-Wong method and learn from each iteration. In the second policy, our objective is to determine a trajectory that expedites the attainment of the final subproblem dual solution. We train and assess these two policies through extensive computational experiments on a network design problem with flow subproblem, confirming that the presence of such learned policies significantly enhances the efficiency of the decomposition process.

Optimization and Control