Luca Ferrarini
Maître de Conférences, LIPN, Université Sorbonne Paris Nord
Embedded Files
I am an Assistant Professor working in the research group AOC at the Université Sorbonne Paris Nord. I obtained my Ph.D. qualification under the supervision of Prof. Stefano Gualandi with a thesis entitled "Polyhedral approach to total matching and total coloring problems". My research interests lie mainly in Polyhedral Combinatorics, Integer Programming, Graph theory, and Machine Learning. I possess a profound enthusiasm for various facets of operations research, both in its theoretical aspects and practical applications across diverse domains. For instance, the intersection of mathematics and music has always intrigued me and I discovered the world of integer programming as a powerful tool to solve and model hard musical problems in mathematical terms. During my Ph.D. studies, I approached the research field of integer programming, with a particular interest in polyhedral approaches and algorithmic point of view. This connection allowed me to develop skills in the fields of discrete optimization and polyhedral theory. My previous work experience at the University of Ecole des Ponts also allows me to broaden my research interest, especially in the field of Machine Learning and Data-Driven Optimization.
RESEARCH INTERESTS
COMBINATORIAL OPTIMIZATION
COMBINATORIAL OPTIMIZATION
Total Coloring and Total Matching Problems.
The aim of the project is to propose exact approaches to tackle the Total Coloring Problem and the Total Matching Problem from a polyhedral perspective.
TILING PROBLEMS IN MUSIC
TILING PROBLEMS IN MUSIC
Tiling Rhytmic canons.
Vuza canons are metric structures that represents melodies without internal repetitions or superposition of voices. These structures has a mathematical counterparts that still yield many open problems. In particular, encoding models allow to discover useful insights on the problem.
MACHINE LEARNING AND Stochastic OPTIMIZATION
MACHINE LEARNING AND Stochastic OPTIMIZATION
Contextual Stochastic Optimization using machine learning pipelines.
Data-driven optimization relies on contextual data and machine learning algorithms to resolve decision problems entangled with uncertain parameters. The objective is to devise efficient algorithms enriched by hybrid pipelines to learn policies that enable us to derive good decisions.
Contacts, web
Page updated
Report abuse