Main research topics

• Markov processes. I develop different activities around the analysis of Markov models from different viewpoints and in different application areas. I am interested in the methodological aspects of Markov chain theory, in the effective analysis of very large models, either numerically or using Monte Carlo techniques. I also work specifically on the transient analysis of Markov models, in particular on looking for closed-forms for the basic distributions. I also develop bounding techniques for queues and networks of queues.
• Dependability analysis. I have a long trajectory around the dependability analysis of complex systems, both in the static case where time is not an explicit model variable, and in the dynamic case where the system is modeled by a stochastic process (a Markovian one, a semi-Markov process, a process with rewards, …)..
• Perceptual Quality and the PSQA technology. I developed the PSQA technology for measuring the Perceptual Quality of applications or services running on the Internet, automatically, accurately and in real-time if necessary. PSQA works both for one-way communications (e.g., in the analysis of a video streaming app) or bi-directional ones (e.g., in the evaluation of an IP phone service). It is based on Supervised Learning tools.
• Performance evaluation of networks. I work on the analysis of the performance of different types of networking systems: access networks in a mobile architecture, optical infrastructures with and without wavelength conversion, peer-to-peer systems. This makes me work in particular on network access, network control, network design.
• Machine Learning. As a side effect of my work on Perceptual Quality assessment, I started research work on Machine Learning tools, often around the concept of Random Neural Network. I have also started work on the use of these types of tools for the analysis of some cognitive human pathologies (namely, the semantic dementia). I also develop methodology in the area (in the training phase when using RNNs for Supervised Learning, in tools for time series prediction). I am also involved in activities around the use of Reinforcement Learning in networking problems.