688 / 2024-05-10 20:26:34

Distributionally robust optimization for pre-disaster facility location problem with 3D printers

Disaster relief,3D printing,Distributionally robust optimization

The continuous advancements in 3D printing technology provide a novel solution for addressing challenges in disaster relief efforts. 3D printers can manufacture essential disaster relief resources using diverse printing materials, effectively meeting the varied demand for assistance in the aftermath of disasters. In this paper, we concentrate on optimizing the costs associated with disaster relief by tackling both the location problem of disaster relief facilities and the allocation predicament of 3D printers and their resources before disasters occur. Given the high unpredictability of natural disasters, we further introduce a two-stage distributionally robust optimization model to handle the uncertainty in demand for various relief resources. The first stage of the model focuses on decisions related to pre-disaster facility location and resource allocation, while the second stage addresses the post disaster rescue activities. Specifically, we propose an ambiguity set for demand uncertainty utilizing the Wasserstein distance and reformulate the two-stage distributionally robust optimization model into a tractable formulation. To solve this problem, we employ a Benders decomposition algorithm with an acceleration strategy. The performance of our proposed model and algorithm is evaluated via a real-world case and large-scale examples. Numerical experiments reveal that our distributionally robust optimization model outperforms the benchmark model across various metrics. Furthermore, we conduct a series of sensitivity analyses and verify the significance of deploying 3D printers before disasters.