Robust facility location in reverse logistics
研究了废木料逆向物流网络中的设施选址问题，建立考虑规模经济和鲁棒性的数学模型，提出基于双层优化的精确求解方法和启发式算法，并在真实与人工数据集上测试性能。
As environmental awareness is becoming increasingly important, alternatives are needed for the traditional forward product flows of supply chains. The field of reverse logistics covers activities that aim to recover resources from their final destination, and acts as the foundation of the efficient backward flow of these materials. Designing the appropriate reverse logistics network for a given field is a crucial problem, as this provides the basis for all operations connected to the resource flow. This paper focuses on design questions in the supply network of waste wood, dealing with its collection and transportation to designated processing facilities. The facility location problem is studied for this use-case, and mathematical models are developed that consider economies of scale and the robustness of the problem. A novel approach based on bilevel optimization is used for computing the exact solutions of the robust problem on smaller instances. A local search and a tabu search method is also introduced for solving problems of realistic sizes. The developed models and methods are tested both on real-life and artificial instance sets in order to assess their performance.