3181. Advances in Optimization Under Uncertainty: From Tractable Multi-Stage Decision Making and Fair Robust Classification to Sustainable Logistics
Invited abstract in session TB-4: EURO Journal on Computational Optimization (EJCO), stream OR Journals.
Tuesday, 10:30-12:00Room: Rupert Beckett LT
Authors (first author is the speaker)
| 1. | Francesca Maggioni
|
| Department of Management, Information and Production Engineering, University of Bergamo |
Abstract
This short talk highlights three research directions within the broader field of optimization under uncertainty, spanning sequential decision-making, machine learning and sustainable logistics.
The first research line addresses the tractability of multi-stage and multi-horizon optimization under uncertainty. We develop bounding and approximation techniques based on scenario grouping, scenario reduction, decomposition methods and constraint sampling. These approaches enable efficient decision-making in complex settings offering theoretical guarantees.
The second focuses on robust and fair support vector machines. We develop models that account for uncertainty in training data through robust and distributionally robust optimization, enhancing classification reliability and fairness to reduce bias with respect to sensitive attributes.
The third direction is dedicated to sustainable logistics, with a focus on fleet sizing, routing, and electric vehicle operations under uncertainty. We propose stochastic programming models to jointly minimize operational costs and environmental impact, accounting for uncertain demand and stochastic energy consumption. Solutions combine decomposition algorithms and tailored heuristics, applied to problems such as Last Mile delivery and electric vehicle routing with threshold-based recharging.
Keywords
- Sustainable Development
- Logistics
Status: accepted
Back to the list of papers