Multi-Agent Communication Systems

This course begins with an overview of various controlled communication systems, establishing a foundation in the principles and challenges of such systems.  Then it introduces the general framework of Markov Decision Processes (MDPs) and their extension to Partially Observable Markov Decision Processes (POMDPs), offering applications of decision-making in uncertain communication environments.

In this lecture the students learn to model communication systems as MDPs or POMDPs, which enables structured analysis and optimization of system performance under different conditions. Building on this, the course introduces practical reinforcement learning techniques tailored for solving/applying (PO)MDPs in communication systems. The focus is initially on single-agent systems and later progresses to multi-agent communication systems, where interactions between agents introduce additional complexity. Students will learn how RL can be extended to manage and optimize these interactions, with a focus on both cooperative and competitive strategies.

Contents

  • Introduction to the relevant classes of communication systems
  • Introduction to MDPs and POMDPs
  • Modelling of communication systems as (PO)MDP
  • Introduction to Reinforcement learning for communication System (PO)MDPs
  • Learning in single-agent communication systems
  • Approaches to multi-agent communication systems (exact, decompositional, approximative)
  • Learning in multi-agent communication systems
  • Cooperative vs competitive communication system modelling

Dates

Summer term

Lecture

  • TBA

Exercise

  • TBA

Materials

Materials will be distributed via Stud.IP

Exam

Written Exam (90 min)

Prüfungs- und Vorlesungsanmeldung


Ihr Dozent

Prof. Dr. Amr Rizk
Professorinnen und Professoren
Adresse
Appelstraße 9a
30167 Hannover
Gebäude
Raum
Prof. Dr. Amr Rizk
Professorinnen und Professoren
Adresse
Appelstraße 9a
30167 Hannover
Gebäude
Raum

Ihre Betreuerinnen und Betreuer

Dr.-Ing. Anam Tahir
Wissenschaftliche Mitarbeiterinnen und Mitarbeiter
Adresse
Appelstraße 9a
30167 Hannover
Gebäude
Raum
Dr.-Ing. Anam Tahir
Wissenschaftliche Mitarbeiterinnen und Mitarbeiter
Adresse
Appelstraße 9a
30167 Hannover
Gebäude
Raum