This project aims at exploring the benefits of recent developments in Machine Learning (ML) in designing better models and solution methods for discrete optimization problems, in particular in the field of Supply Chain and Logistics. Our research will be focused on two directions. First, we will try to use ML techniques to incorporate demand learning into optimization problems. This will lead to increased responsiveness of the supply chain processes modeled. The joint optimization of demand learning and of the supply chain process will increase the scale of the problem, while maintaining its linearity. One of our goals is to design efficient algorithms for the joint problem.
Second, we will investigate methods to learn the structure of optimal solutions of computationally tractable problems and use this structure to solve efficiently large-scale supply chain optimization problems, such as vehicle routing or delivery problems.
This project will focus on supply chain, an essential part of any large company. Hopefully, the techniques developed in this project will lead to a higher efficiency and responsiveness of supply chains.
