Facility location decisions play a critical role in transportation planning. In this talk, we study a capacitated fixed-charge location-transportation problem in which, while the location and capacity of each facility need to be determined immediately, the determination of final production and distribution of products can be delayed until actual orders are received. In contexts where little is known about future demand, robust optimization becomes a natural method to employ in order to identify meaningful decisions. Unfortunately, such problems are computationally intractable. To overcome this difficulty, we propose a set of conservative approximations to the problem that each exploits to a different extent the idea of reducing the flexibility of the delayed decisions. Each approximation will have the potential to reach a different level of compromise between efficiency of resolution and quality of the solution. We will also demonstrate that full flexibility is often unnecessary to reach nearly, or even exact, optimal robust locations and capacities for the facilities. We illustrate our findings with a numerical study where we evaluate the effect of the amount of uncertainty on the performance and structure of each approximate solution that can be obtained.
Ce séminaire s'adresse seulement aux étudiants et membres du GERAD. Nous vous remercions de confirmer votre présence. Des pizzas et des breuvages seront servis aux participants ou vous pouvez apporter votre lunch.