Home
Research
Brief cv
CV (on LinkedIn)
DBLP
Google scholar

 

 

 

 

 

Hi, thanks a lot for your interest in my work. My name is Michail Fasoulakis and I come from Greece and, in particular, from the beautiful island of Crete.

I am currently an Assistant Professor (Lecturer) in (Theoretical) Computer Science at Royal Holloway, University of London, UK, and an Affiliated/Corresponding Scientist of Abroad (Honorary Title) at the ICS-FORTH.

 

I am an (Applied) Mathematician and (Theoretical) Computer Scientist with a background and interest in (Computational) Economics/Game theory/Operations Research and Computer Engineering.

My academic background is a mix of (theoretical) Computer Science, (Applied) Mathematics, Computer Engineering and (Computational) Microeconomics/Operations Research, having almost 22 years of Education/Research experience.

 

My research interests lie in basic and fundamental topics (in abstract structures) for Algorithms, Game theory/Decision theory, Optimization, Mathematics of Information (Information theory), and their (theoretical) applications, especially, in AI/ML, Communications and Signal Processing, Economics and Operations Research.

My main goal is to try to contribute in research introducing new mathematical/theory concepts, or solving analytically/algorithmically abstract problems in these research areas that can be potentially applied in real applications. Furthermore, I am interested in solving problems in specific applications of these areas.

 

More specifically, with my coauthors, I have intensively worked in issues of game theory and their algorithmic view, like the computation of (approximate) Nash equilibria. This line of work had as a result the following selected contributions:

    "In this publication we give a polynomial-time algorithm for computing (1/3+δ)-Nash equilibria, giving the best bound for approximate Nash equilibria at the moment of publication, an improvement for a fundamental problem in algorithmic game theory open for 15 years."


    "In this publication we give a polynomial-time algorithm for computing (1/2+δ)-well-supported Nash equilibria, giving the best bound for approximate well-supported Nash equilibria at the moment of publication, an improvement for a fundamental problem in algorithmic game theory open for 7 years."


    M. Fasoulakis, E. Markakis, Y. Pantazis, and C. Varsos. "Forward Looking Best-Response Multiplicative Weights Updated Methods for bilinear zero-sum games". AISTATS 2022.

    "In this publication we give a learning algorithm for last iterate convergence in zero-sum bimatrix games (a fundamental class of bimatrix games in game theory)."


    "In this publication we give a new algorithm for computing (approximate) Nash equilibria in zero-sum bimatrix games."


Another recent Research highlight is the following work:

    M. Fasoulakis, K. Varsos, A. Traganitis. "Revisit the Arimoto-Blahut algorithm: New analysis with approximation". Preprint at Arxiv.

    "In this publication, we give an improved approximation analysis that gives approximation guarantees with inverse exponential rate of convergence, that implies better upper bound for the total complexity for the Arimoto-Blahut algorithm, an almost 50 years algorithm and one of the fundamental algorithms in Information theory for computing the Shannon's Capacity of a Channel."


 

Michail Fasoulakis, Ph.D.


Contact info:

at x@rhul.ac.uk,

{where x = Michail.Fasoulakis},

at y@ics.forth.gr,

{where y = mfasoul}.

Last update of the webpage: 10/03/2025.