Research

Network cross-layer optimization

Cross-layer optimization breaks the boundaries between layers of a communication model by consider joint optimization of multiple aspects at once. In particular, topics investigated relate to joint routing and scheduling, channel assignment, node positioning, and resource allocation. Such a complex task often requires resorting to meta-heuristic techniques such as genetic algorithms, multi-armed bandits, or other machine learning techniques.

Age-of-Information

Especially useful for remote sensing scenarios, Age-of-information (AoI) quantifies freshness of data by characterizing it as the time elapsed since the latest useful update. It involves closed-form computations from classic dynamic programming and queueing theory, also with possible repercussions on Game theory and linear programming.

Transmission protocol modelling via stochastic approaches

Markov models allow for rigorous analytical characterization of many transmission techniques. In particular, retransmission-based techniques such as ARQ or hybrid ARQ have been analyzed, considering different aspects such as delay characterization, performance of multimedia transmissions, resulting AoI.

Game theory

Game theory is the study of multi-person multi-objective problems. Yes, albeit the name is cool, it has little to do with video games (even though, strictly speaking even multi-player games belong to the category of multi-agent optimization). It allows precise representation of distributed management scenarios, which is extremely useful when analyzing network scenarios. Notably, it is not necessarily restricted to conflict situations, but it can also describe contexts where multiple players act without any coordination and therefore the management can be improved.