Reusable Benders cuts for bi-objective mixed integer linear programming
Sprache des Vortragstitels:
Englisch
Original Tagungtitel:
International Conference on Operations Research (OR) 2023
Sprache des Tagungstitel:
Englisch
Original Kurzfassung:
We investigate bi-objective mixed integer linear programming problems to which Benders decomposition can be applied. Benders decomposition is a widely used technique for exploiting problem structures that arise in real-world applications,
including supply chain network design problems. Typically, the Benders master problem concerns the strategic decisions such as the facility location decisions while the Benders subproblem deals with the operational decisions, such as commodity flow. Our study focuses on a sustainable
supply chain network design problem that aims to minimise both economic and environmental objectives. To address the two objective nature of the problem, we employ the weighted-sum method. The following issue then is that we repeatedly have to solve similar problems with different weight
values. To tackle this, we propose a simple way to reuse the generated Benders cuts in subsequent iterations of the weighted sum method. The underlying idea is that if the weights are similar, the Benders master problem would not require too many new Benders cuts to be generated,
enabling us to start the Benders decomposition for the master problem with some pre-generated Benders cuts.