Algorithmic game theory: Difference between revisions

Revision as of 18:10, 19 December 2020

Latest revision as of 17:09, 9 August 2025

(18 intermediate revisions by 17 users not shown)

Line 1:

{{Short description|Study of algorithms in strategic environments}}

{{essay-like|date=August 2013}}

'''Algorithmic game theory''' ('''AGT''') is an areainterdisciplinary field inat the intersection of [[game theory]] and [[computer science]], withfocused the objective ofon understanding and design ofdesigning algorithms infor [[Strategyenvironments (gamewhere multiple theory)|strategic]] environmentsagents interact. This research area combines computational thinking with economic principles to address challenges that emerge when algorithmic inputs come from self-interested participants.

In traditional [[algorithm design]], inputs are assumed to be fixed and reliable. However, in many real-world applications—such as [[Electronic auction|online auctions]], [[internet routing]], [[digital advertising]], and [[resource allocation]] systems—inputs are provided by multiple independent agents who may strategically misreport information to manipulate outcomes in their favor. AGT provides frameworks to analyze and design systems that remain effective despite such strategic behavior.

Typically, in Algorithmic Game Theory problems, the input to a given algorithm is distributed among many players who have a personal interest in the output. In those situations, the [[agent (economics)|agent]]s might not report the input truthfully because of their own personal interests. We can see Algorithmic Game Theory from two perspectives:

The field can be approached from two complementary perspectives:

* ''Analysis'': look at the current implemented algorithms and analyze them using Game Theory tools: calculate and prove properties on their [[Nash equilibria]], [[price of anarchy]], best-response dynamics ...

* ''Design'': design games that have both good game-theoretical and algorithmic properties. This area is called [[algorithmic mechanism design]].

* ''Analysis'': Evaluating existing algorithms and systems through game-theoretic tools to understand their strategic properties. This includes calculating and proving properties of [[Nash equilibria]] (stable states where no participant can benefit by changing only their own strategy), measuring [[price of anarchy]] (efficiency loss due to selfish behavior), and analyzing best-response dynamics (how systems evolve when players sequentially optimize their strategies).

On top of the usual requirements in classical algorithm design, say ''polynomial-time running time'', ''good approximation ratio'', ... the designer must also care about incentive constraints.

* ''Design'': Creating mechanisms and algorithms with both desirable computational properties and game-theoretic robustness. This sub-field, known as [[algorithmic mechanism design]], develops systems that incentivize truthful behavior while maintaining computational efficiency.

Algorithm designers in this ___domain must satisfy traditional algorithmic requirements (such as ''[[polynomial-time]] running time'' and ''good approximation ratio'') while simultaneously addressing incentive constraints that ensure participants act according to the system's intended design.

==History==

===Nisan-Ronen: a new framework for studying algorithms===

In 1999, the seminal paper of [[Noam Nisan|Nisan]] and Ronen[[Amir Ronen]]<ref name="nr">{{citation

| last1 = Nisan | first1 = Noam | author1-link = Noam Nisan

| last2 = Ronen | first2 = Amir | author2-link = Amir Ronen

| contribution = Algorithmic mechanism design

| doi = 10.1145/301250.301287

| pages = 129–140

| title = Proceedings of the 31st ACM Symposium on Theory of Computing (STOC '99)

| year = 1999| isbn = 978-1581130676 | s2cid = 8316937 | doi-access = free}}</ref> drew the attention of the Theoretical Computer Science community to designing algorithms for selfish (strategic) users. As they claim in the abstract:

{{QuoteBlockquote|We consider algorithmic problems in a distributed setting where the participants cannot be assumed to follow the algorithm but rather their own self-interest. As such participants, termed agents, are capable of manipulating the algorithm, the algorithm designer should ensure in advance that the agents’ interests are best served by behaving correctly.

Following notions from the field of mechanism design, we suggest a framework for studying such algorithms. In this model the algorithmic solution is adorned with payments to the participants and is termed a mechanism. The payments should be carefully chosen as to motivate all participants to act as the algorithm designer wishes. We apply the standard tools of mechanism design to algorithmic problems and in particular to the shortest path problem.}}

Line 39 ⟶ 41:

{{main|Price of Anarchy}}

The other two papers cited in the 2012 Gödel Prize for fundamental contributions to Algorithmic Game Theory introduced and developed the concept of "Price of Anarchy".

In their 1999 paper "Worst-case Equilibria",<ref name = "kp">{{Cite journal|title=Worst-case Equilibria|first1=Elias|last1=Koutsoupias|first2=Christos|last2=Papadimitriou|journal=Computer Science Review|volume=3|issue=2|date=May 2009|pages=65–69|url=http://www.cs.berkeley.edu/~christos/nash.ps|doi=10.1016/j.cosrev.2009.04.003|access-date=2018-01-08|archive-url=https://web.archive.org/web/20160313023635/http://www.cs.berkeley.edu/~christos/nash.ps|archive-date=2016-03-13|url-status=dead}}</ref> Koutsoupias and [[Christos Papadimitriou|Papadimitriou]] proposed a new measure of the degradation of system efficiency due to the selfish behavior of its agents: the ratio of between system efficiency at an optimal configuration, and its efficiency at the worst Nash equilibrium. (The term "Price of Anarchy" only appeared a couple of years later.<ref>{{citation

(The term "Price of Anarchy" only appeared a couple of years later.<ref>{{citation

| last = Papadimitriou | first = Christos | author-link = Christos Papadimitriou

| contribution = Algorithms, games, and the Internet

Line 67 ⟶ 68:

The existence of an equilibrium in a game is typically established using non-constructive [[fixed point theorem]]s. There are no efficient algorithms known for computing [[Nash equilibrium|Nash equilibria]]. The problem is complete for the [[complexity class]] [[PPAD]] even in 2-player games.<ref name="chen2005">*{{Cite conference|first1=Xi|last1=Chen|first2=Xiaotie|last2=Deng|title=Settling the complexity of two-player Nash equilibrium|conference=Proc. 47th Symp. Foundations of Computer Science|year=2006|pages=261–271|doi=10.1109/FOCS.2006.69|id={{ECCC|2005|05|140}}}}.</ref> In contrast, [[correlated equilibrium|correlated equilibria]] can be computed efficiently using linear programming,<ref>{{cite journal |first1=Christos H. |last1=Papadimitriou |first2=Tim |last2=Roughgarden |title=Computing correlated equilibria in multi-player games |journal=J. ACM |volume=55 |issue=3 |pages=14:1–14:29 |year=2008 |doi=10.1145/1379759.1379762|citeseerx=10.1.1.335.2634 |s2cid=53224027 }}</ref> as well as learned via no-regret strategies.<ref>{{cite journal |last1=Foster |first1=Dean P. |first2=Rakesh V. |last2=Vohra |title=Calibrated Learning and Correlated Equilibrium |journal=Games and Economic Behavior |year=1996 |url=https://repository.upenn.edu/cgi/viewcontent.cgi?article=1008&context=statistics_papers}}</ref>

=== [[Computational social choice]]===

{{Main|Computational social choice}}

Computational social choice studies computational aspects of ''social choice'', the aggregation of individual agents' preferences. Examples include algorithms and computational complexity of voting rules and coalition formation.<ref>{{citation

Line 103 ⟶ 104:

Algorithmic Game Theory papers are often also published in Game Theory journals such as [[Games and Economic Behavior|GEB]],<ref>

{{citation

| last1 = Chawla | first1 = Shuchi | author1-link = Shuchi Chawla

| last2 = Fleischer | first2 = Lisa

| last3 = Hartline | first3 = Jason

Line 119 ⟶ 120:

*[[Auction Theory]]

*[[Computational social choice]]

*[[Gamification]]

*[[Load balancing (computing)]]

*[[Mechanism design]]

Line 140 ⟶ 142:

==External links==

*[httphttps://gambit.sourceforge.net/ gambit.sourceforge.net] - a library of game theory software and tools for the construction and analysis of finite extensive and strategic games.

*[http://gamut.stanford.edu/ gamut.stanford.edu] - a suite of game generators designated for testing game-theoretic algorithms.

{{Authority control}}

{{DEFAULTSORT:Algorithmic Game Theory}}

[[Category:Game theory|+]]

[[Category:Theory of computation]]

[[Category:Algorithms]]