Algorithmic Game Theory (Winter 2024/25)
Topical
- Logbook, with details about schedule and material.
- There will be no lectures on December 16+17, 2024. The next lecture is on Monday, Jan 6, 2025.
- Exercise sessions will continue as usual in the week 16-20 Dec.
- Solutions to sheet 8 shall be submitted on Dec 17, and sheet 9 is issued (for submission on Jan 7, 2025).
- The last lecture will be on January 20, 2025.
Organizational
- Lecturers: Prof. Dr. Martin Hoefer
- Exercises: Sukanya Pandey, Rilind Sahitaj
- RWTH Online: course page, exercise page
- Room in RWTH Moodle
- Lectures:
Mondays, 10:30 - 12:00h in AH III (Informatikzentrum)
Tuedays, 10:30 - 12:00h in AH II (Informatikzentrum) - Exercises: TBD
- First Exam: Feb 20, 2025. (written exam)
- Second Exam: Mar 18, 2025. (written exam)
Material
Slides and Notes
| Chapter | Updates | Lectures | |
|---|---|---|---|
| Organizational | Slides | 08.10.2024 | 1 |
| Strategic Games and Nash Equilibrium | Slides | 07.10.2024 | 1-3 |
| Pure Nash Equilibria | Slides | 09.10.2024 | 3-7 |
| Learning and Correlated Equilibria | Slides | 05.11.2024 | 7-9 |
| Prices of Anarchy and Stability | Slides | 19.11.2024 | 10-12 |
| Designing Incentive-Compatible Mechanisms | Slides | 09.12.2024 | 13-16 |
| Social Choice | Slides | 09.12.2024 | 17-19 |
Lecture Notes
German lecture notes from a previous version of this course are available here.Exercise Sheets
Weekly exercise sheets will be published here. Solutions must be composed by groups of (initially) 3 students.
Your solutions must be submitted as a single PDF file via Moodle.
You must score at least 50% of the total number of points to be admitted to the exam. If you score at least 75%, you can obtain one grading step bonus for the exam. To receive the bonus, you must pass the exam, and at least one solution must be presented during an exercise session.
Content
The course provides an introduction to theoretical and algorithmic foundations of computer systems that involve strategic and economic interaction of rational agents. These systems arise frequently in modern computer networks -- service providers strive to route packets as quickly or cheap as possible, in cloud computing the resources (such as computing time or memory) are shared, rented or sold, advertisers want to place their ads as prominently as possible and pay as little as possible, etc. The business model of many companies relies on trade and marketing in computational markets on the Internet.
In algorithmic game theory we design and analyze algorithms for systems with interaction of many rational agents. These algorithms search for optimal strategies for single users, or they try to optimize performance for the system while addressing strategic behavior of users. The goals are a characterization of incentives, as well as provable bounds on running time and solution quality for optimization algorithms. In the course, we will introduce basic ideas from game theory and combine them with techniques from approximation algorithms, distributed computing, and complexity theory.
Literature
Directly related to the course material:
- [EK] Easley, Kleinberg. Networks, Crowds, and Markets. Cambridge University Press, 2010.
Access here. - [NRTV] Nisan, Roughgarden Tardos, Vazirani. Algorithmic Game Theory. Cambridge University Press, 2007.
Non-printable PDF here. - [R] Roughgarden. Twenty Lectures in Algorithmic Game Theory. Cambridge University Press, 2016.
- [RBLR] Rothe, Baumeister, Lindner, Rothe. Einführung in Computational Social Choice. Spektrum Verlag, 2012. Access with Goethe University license here.
Many textbooks cover background and context in game theory, e.g.,
- [M] Myerson. Game Theory: Analysis of Conflict. Harvard University Press, 1991.
- [O] Owen. Game Theory. Academic Press, 2001.
- [OR] Osborne, Runbinstein. A Course in Game Theory. MIT Press, 1995.
- etc...