Journées de l'optimisation 2022

HEC Montréal, Québec, Canada, 16 — 18 mai 2022

Horaire Auteurs Mon horaire

WA9 - Applications of distributionally robust optimization

18 mai 2022 10h30 – 12h10

Salle: Groupe Cholette (jaune)

Présidée par Chun Peng

4 présentations

  • 10h30 - 10h55

    Network Flow Models for Two-Stage Robust Binary Optimization

    • Ian Yihang Zhu, prés., University of Toronto
    • Merve Bodur, University of Toronto
    • Timothy C.Y. Chan, University of Toronto

    In this talk, we examine two-stage robust binary optimization (RBO) problems with objective uncertainty. We describe a new set of reformulation techniques inspired by decision diagrams which enable the transformation of the infinite-dimensional two-stage RBO problem into a single constrained network flow model. We show that the network flow model is easy to implement and performs well computationally compared to existing solution methods.

  • 10h55 - 11h20

    Robust Integration of Electric Vehicles Charging Load in Smart Grid’s Capacity Expansion Planning

    • Sajad Sani, prés., HEC
    • Erick Delage, GERAD, HEC Montréal
    • Olivier Bahn, HEC Montréal
    • Rinel Foguen Tchuendom, Ecole Polytechnique à Montréal

    Battery charging of electric vehicles (EVs) needs to be properly coordinated by electricity producers to maintain network reliability. We propose a robust approach to model the interaction between a large fleet of EV users and utilities in a long-term generation expansion planning problem. In doing so, we employ a robust multi-period adjustable generation expansion planning problem, called R-ETEM, in which demand responses of EV users are uncertain. Then, we employ a linear-quadratic game to simulate the average charging behavior of EV users. The two models are coupled through a dynamic price signal broadcasted by the utility. Mean-field game theory is used to solve the linear-quadratic game model. Finally, we develop a new coupling algorithm between R-ETEM and the linear-quadratic game with the purpose of adjusting in R-ETEM the uncertainty level of EV demand responses.

  • 11h20 - 11h45

    Distributional Robustness and Inequity Mitigation in Disaster Preparedness of Humanitarian Operations

    • Hongming Li, prés., HEC Montréal
    • Erick Delage, GERAD, HEC Montréal
    • Ning Zhu, Tianjin University
    • Michael Pinedo, New York University
    • Shoufeng Ma, Tianjin University

    In this paper, we study a predisaster relief network design problem with uncertain demands. The aim is to determine the prepositioning and reallocation of relief supplies. Motivated by the call of the International Federation of Red Cross and Red Crescent Societies (IFRC) to leave no one behind, we consider three important practical aspects of humanitarian operations: shortages, equity, and uncertainty. Numerical studies on the 2010 Yushu earthquake positively demonstrate the value of our methodology in alleviating geographical inequities and reducing shortages. Our case study also provides several other interesting insights that may be useful for humanitarian organizations for disaster preparedness.

  • 11h45 - 12h10

    Distributionally robust optimization for multi-item lot sizing problem under yield uncertainty

    • Paula Metzker Soares, prés.,
    • Simon Thevenin, IMT Atlantique - LS2N
    • Yossiri Adulyasak, HEC Montréal
    • Alexandre Dolgui, IMT Atlantique - LS2N

    This research investigates a distributionally robust optimization approach to the lot-sizing problem with backorder and yield uncertainty under event-wise ambiguity sets. We rely on the Moment-based, Wasserstein and K-Means clustering ambiguity sets to represent the yield distribution. We also consider static and static-dynamic decision strategies to calculate a solution. In this talk, we will present the performance of different sets of ambiguity to determine a production plan that is satisfactorily robust to changes in the environment. The numerical experiment shows that the model remains tractable for all ambiguity sets considered and the production plans obtained remain efficient for different strategies and decision contexts.