Adaptive Model Predictive Control in the Inventory Control Problem

The present thesis aims to develop and to analyze methods and algorithms to control uncertain systems. The methods developed are based on control and optimization theory. Model Predictive Control, Large Deviations, Approximate Dynamic Programming and Robust Optimization are applied, extended and combined to face the challenges presented in the application of such methods in real-world problems. In particular, the algorithms developed have been applied to supply chain systems. A supply chain is composed by several business units working together to match the market demand of a product. Despite several economic and cultural changes, (i.e. low production cost, international outsourcing...) the main goal of a supply chain networks is to procure raw materials and transform them into final products. This involves the automation of several processes: material and information flows, and relationships between supplier and customers. To reach these objective the management of a supply chain has to take several decisions and to supervise and to control several facilities.

Autorentext

He is a Quant Strategist at JP Morgan Chase & Co. He also is Visiting Professor of Operations Research at the Free University of Bozen. Previously, he has been a researcher at University of Cambridge. He holds a PhD in Artificial Intelligence from the Universita' Politecnica delle Marche. He has been a Visiting Scholar at Boston University.