Cooperative and heterogeneous multi-agent learning for 6G network orchestration

Published : 18 January 2022

In beyond 5G/6G networks , it is imperative to easily deploy and manage a private/ad-hoc network of mobile users such as a fleet of vehicles or drones. The objective of this thesis is to define strategies and associated protocols (control and resource allocation) to self-organize “mesh” networks of mobile users.

The research questions are: (i) How to manage a cooperative multi-agent system for the orchestration and self-organization of a 6G network? (ii) How to orchestrate a distributed multi-objective network? (iii) Are the multi-agent approach and network reconfiguration compatible with the dynamics of the environment?

While existing studies focus on problems aiming at optimizing a single objective function with homogeneous agents, we are interested in local/distributed multi-agent cooperative learning between heterogeneous users/moving agents (with different optimization functions).

The first step of this thesis will be to optimize heterogeneous multi-objective functions for a 6G network with a central orchestrator. The second step of this thesis will concern cooperative heterogeneous multi-agent systems and interactions between agents (concurrent learning, team learning, …) to jointly solve tasks and maximize utility. The last step of this thesis will concern a Hybrid approach (Centralized and Distributed)

More information