Uso de cookies

En las páginas web de la Universidad Carlos III de Madrid utilizamos cookies propias y de terceros para mejorar nuestros servicios mediante el análisis de sus hábitos de navegación. Al continuar con la navegación, entendemos que se acepta nuestra política de cookies. "Normas de uso"

Roberto Battiti


Roberto Battiti - University of Trento, Italy

Roberto was born in Trento (a small Italian town with an intriguing history, barycenter of Europe in the Alps) and received the Laurea in Theoretical Physics from the University of Trento in 1985. Then he moved to the USA, where he received the Ph. D. degree from the California Institute of Technology (Caltech) in 1990. As full professor of Computer Science he organizes research initiatives in the area of reactive search optimization (RSO) and learning and intelligent optimization (LION) heuristics. His passion is to use data to build flexible models and extract actionable knowledge (machine learning), to exploit knowledge to automate the discovery of improving solutions (intelligent optimization), to connect insight to decisions and actions ("prescriptive analytics"). Roberto is the director of the LION lab: machine Learning and Intelligent OptimizatioN for prescriptive analytics, author of more than 100 highly-cited publications, with a rich experience in startup activities. He is a Fellow of the IEEE.

Complete and updated details about his work can be found in his webpage:



The research of Dr. Battiti spans across multiple disciplines, in particular machine learning and intelligent optimization.

The first direction he expects his work to take is to explore machine learning and optimization methods for a possible application in the cross-layer optimization of the overall communications network. Dr. Battiti will work in collaboration with experts in networking science from UC3M and IMDEA Networks to identify specific areas of investigation in this area.

In the first place, machine learning from usage data related to demand, energy consumption and quality of service in different conditions will be performed. This is critical to develop dynamic models. Secondly, when the first models are available, a fast optimization process can act on the models, for example, to relocate the capacity of the backhaul/fronthaul to a different link in case the current situation requires it.  Finally, after initialization, the machine learning plus the optimization process run continuously, so that the dynamic models can be updated to reflect the current situation of the network and automated reallocation schemes can improve the overall quality measures. The dynamic capacity reconfiguration should happen in real-time, aiming at improving critical measures, energy efficiency and quality of service. Research and possible startup initiatives in the area of internet services will also be investigated.