Alternative formulations and solution approaches for distribution network design with seasonality

Official URL: https://risc01.sabanciuniv.edu/record=b2553609 _(Table of contents)

In this study, we consider a single-commodity distribution network design problem, which takes seasonal variations in the demand into account. We consider a three-echelon supply chain network design over a planning horizon, consisting of four seasons; products are delivered to outlets from a distribution center through regional depots. We develop alternative mathematical models that have different levels of flexibility while responding to seasonal demand. The problem formulations incorporate decisions related to locations of regional depots, amount of transportation from distribution center to regional depots, and routes used for delivery from regional depots to outlets while the objective function minimizes the total cost due to opening and operating regional depots as well as transportation-related costs. To solve the resulting problems, we first propose a linear-programming-based constructive heuristic approach. Alternatively, we adapt the local branching algorithm to all three models with variations on branching of different binary decision variables. In order to evaluate the efficiency and effectiveness of the proposed heuristics, we solve instances of four sets of problems varying in terms of the problem size. We also evaluate the effect of the truck size used in delivery to outlets on the problem difficulty and also its impact on the solution quality. The results show that the local branching algorithm has mostly demonstrated a better performance in terms of solution quality and computational efficiency compared to other approach