stage + thèse sur les jeux évolutionnaires (Avignon et Montpellier)

The overall objective of EAGPLAN project, headed by Orange labs, is to design and implement efficient algorithms for computing optimal solutions that will enable to derive recommendations on the best actions that a telecommunication operator can undertake for multiperiod network planning in competitive landscape, considering the actions of the other operators. The project builds on research carried out in previous works that addressed this problem through the lens of extensive-form Games. The findings highlighted the effectiveness of combining tools and methodologies from Graph Theory, Mathematical Programming and Dynamic Programming to compute the outcomes of pure Nash equilibria in extensive-form games. Building upon these promising perspectives, we will propose algorithms to 1) tackle challenging problems arising from multiperiod deployment and decommissioning of network technologies in a competitive environment, while also accounting for uncertainties related to competitors, such as their payoffs values or order of actions and 2) provide solutions that will be used by an operator to derive strategic planning of the actions (for example deployment or decommissioning) necessary to develop its network. We aim at integrating the resulting algorithms in an open source demonstrator for solving game theory problems. This demonstrator will also provide a realistic and large size extensive-form game instances generator that can be used for benchmarking and validation purpose in the scientific community.

Details of the internsips and PhD topics available here: https://seafile.lirmm.fr/f/44a398186a824c90b4a5/?dl=1