PhD "A game-theoretic approach to efficient and sustainable cloud computing systems." M/F

about the role

Your role is to carry out a PhD on “a game-theoretic approach to efficient and sustainable cloud computing”.

• Context and problem

Human societies and technologies must evolve under the pressure of various factors such as natural resources depletion, limited availability of energy or the growth of the population and its needs (including in the digital domain where demand is strongly increasing).

In this context, cloud computing solutions are large-scale systems typically distributed and with high energy and resource consumptions. These systems must be modified or redesigned to be more efficient and more sustainable, explicitly taking into account constraints on resources or energy and their usages.

• Scientific objectives 

This thesis aims at using game theory to analyze and propose solutions to the previous problem. However, existing equilibrium concepts with constraints [I], such as the generalized Nash equilibria [II][III] or (more recently) constrained correlated equilibria [IV] (combining correlation, stability and constraints), are not necessarily suitable for dynamic games, do not allow for some classes of constraints or must be adapted to more general assumptions on the players’ information. Furthermore, these solutions are typically not suitable for implementation on large-scale systems (such as networks and clouds) where distributed architectures must be considered.

The main challenges to address are

• model the problem as a game with constraints
• assess the relevance of constrained correlated equilibria & related developments or define and analyze a suitable solution concept
• propose algorithmic solutions to compute the equilibria 
• assess the performances of the solution.

References

1. Kulkarni, Games and teams with shared constraints, Phil. Trans. R. Soc., 2017.
2. Rosen, Existence and uniqueness of equilibrium points for concave n-person games, Econometrica, 1965. 
3. Facchinei et al., Generalized Nash equilibrium problems. 4OR, Sep 2007.
4. Aumann, Correlated Equilibrium as an Expression of Bayesian Rationality. Econometrica, 1987.