2607. Consensus, upgrade and skeleton clustering for stochastic programming
Invited abstract in session MB-31: Machine Learning for Optimization under uncertainty 1, stream Stochastic and Robust optimization.
Monday, 10:30-12:00Room: Maurice Keyworth 1.06
Authors (first author is the speaker)
| 1. | Janosch Ortmann
|
| GERAD, CRM and UQAM | |
| 2. | Samah Abdalrhaman
|
| AOTI, UQAM | |
| 3. | Walter Rei
|
| CIRRELT and UQAM |
Abstract
In this talk I will discuss two applications from freight transportation and logistics: a collective food network and a stochastic network design problem.
These problems can be modelled as two-stage stochastic optimisation problems, but the number of scenarios required to realistically model the uncertainty typically makes the optimization problem numerically intractable. I will show how to use problem-based clustering to extract information about the stochastic program from the single-scenario solutions.
We introduce new distance matrices on the scenario set that are based on the idea of giving the decision maker some flexibility to change the first-stage decisions after the uncertainty has been revealed. In particular, we integrate the ideas of upgrade and skeleton solutions.
We then apply clustering on the scenario set equipped with these distances. Preliminary result indicate promising bounds and reveal interesting new structures on the scenario set.
Keywords
- Machine Learning
- Stochastic Optimization
- Analytics and Data Science
Status: accepted
Back to the list of papers