Heterogeneous learning in zero-sum stochastic games with incomplete information

Quanyan Zhu, Tembine Hamidou, Tamer Başar

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Learning algorithms are essential for the applications of game theory in a networking environment. In dynamic and decentralized settings where the traffic, topology and channel states may vary over time and the communication between agents is impractical, it is important to formulate and study games of incomplete information and fully distributed learning algorithms which for each agent requires a minimal amount of information regarding the remaining agents. In this paper, we address this major challenge and introduce heterogeneous learning schemes in which each agent adopts a distinct learning pattern in the context of games with incomplete information. We use stochastic approximation techniques to show that the heterogeneous learning schemes can be studied in terms of their deterministic ordinary differential equation (ODE) counterparts. Depending on the learning rates of the players, these ODEs could be different from the standard replicator dynamics, (myopic) best response (BR) dynamics, logit dynamics, and fictitious play dynamics. We apply the results to a class of security games in which the attacker and the defender adopt different learning schemes due to differences in their rationality levels and the information they acquire.

Original languageEnglish (US)
Title of host publication2010 49th IEEE Conference on Decision and Control, CDC 2010
Pages219-224
Number of pages6
DOIs
StatePublished - 2010
Event2010 49th IEEE Conference on Decision and Control, CDC 2010 - Atlanta, GA, United States
Duration: Dec 15 2010Dec 17 2010

Other

Other2010 49th IEEE Conference on Decision and Control, CDC 2010
CountryUnited States
CityAtlanta, GA
Period12/15/1012/17/10

Fingerprint

Stochastic Games
Zero sum game
Incomplete Information
Game
Learning algorithms
Learning Algorithm
Fictitious Play
Replicator Dynamics
Logit
Learning Rate
Stochastic Approximation
Game theory
Game Theory
Rationality
Distributed Algorithms
Dynamic Response
Networking
Parallel algorithms
Ordinary differential equations
Decentralized

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Modeling and Simulation
  • Control and Optimization

Cite this

Zhu, Q., Hamidou, T., & Başar, T. (2010). Heterogeneous learning in zero-sum stochastic games with incomplete information. In 2010 49th IEEE Conference on Decision and Control, CDC 2010 (pp. 219-224). [5718053] https://doi.org/10.1109/CDC.2010.5718053

Heterogeneous learning in zero-sum stochastic games with incomplete information. / Zhu, Quanyan; Hamidou, Tembine; Başar, Tamer.

2010 49th IEEE Conference on Decision and Control, CDC 2010. 2010. p. 219-224 5718053.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Zhu, Q, Hamidou, T & Başar, T 2010, Heterogeneous learning in zero-sum stochastic games with incomplete information. in 2010 49th IEEE Conference on Decision and Control, CDC 2010., 5718053, pp. 219-224, 2010 49th IEEE Conference on Decision and Control, CDC 2010, Atlanta, GA, United States, 12/15/10. https://doi.org/10.1109/CDC.2010.5718053
Zhu Q, Hamidou T, Başar T. Heterogeneous learning in zero-sum stochastic games with incomplete information. In 2010 49th IEEE Conference on Decision and Control, CDC 2010. 2010. p. 219-224. 5718053 https://doi.org/10.1109/CDC.2010.5718053
Zhu, Quanyan ; Hamidou, Tembine ; Başar, Tamer. / Heterogeneous learning in zero-sum stochastic games with incomplete information. 2010 49th IEEE Conference on Decision and Control, CDC 2010. 2010. pp. 219-224
@inproceedings{4701f7f1bea64013bae4a0bb30b7593f,
title = "Heterogeneous learning in zero-sum stochastic games with incomplete information",
abstract = "Learning algorithms are essential for the applications of game theory in a networking environment. In dynamic and decentralized settings where the traffic, topology and channel states may vary over time and the communication between agents is impractical, it is important to formulate and study games of incomplete information and fully distributed learning algorithms which for each agent requires a minimal amount of information regarding the remaining agents. In this paper, we address this major challenge and introduce heterogeneous learning schemes in which each agent adopts a distinct learning pattern in the context of games with incomplete information. We use stochastic approximation techniques to show that the heterogeneous learning schemes can be studied in terms of their deterministic ordinary differential equation (ODE) counterparts. Depending on the learning rates of the players, these ODEs could be different from the standard replicator dynamics, (myopic) best response (BR) dynamics, logit dynamics, and fictitious play dynamics. We apply the results to a class of security games in which the attacker and the defender adopt different learning schemes due to differences in their rationality levels and the information they acquire.",
author = "Quanyan Zhu and Tembine Hamidou and Tamer Başar",
year = "2010",
doi = "10.1109/CDC.2010.5718053",
language = "English (US)",
isbn = "9781424477456",
pages = "219--224",
booktitle = "2010 49th IEEE Conference on Decision and Control, CDC 2010",

}

TY - GEN

T1 - Heterogeneous learning in zero-sum stochastic games with incomplete information

AU - Zhu, Quanyan

AU - Hamidou, Tembine

AU - Başar, Tamer

PY - 2010

Y1 - 2010

N2 - Learning algorithms are essential for the applications of game theory in a networking environment. In dynamic and decentralized settings where the traffic, topology and channel states may vary over time and the communication between agents is impractical, it is important to formulate and study games of incomplete information and fully distributed learning algorithms which for each agent requires a minimal amount of information regarding the remaining agents. In this paper, we address this major challenge and introduce heterogeneous learning schemes in which each agent adopts a distinct learning pattern in the context of games with incomplete information. We use stochastic approximation techniques to show that the heterogeneous learning schemes can be studied in terms of their deterministic ordinary differential equation (ODE) counterparts. Depending on the learning rates of the players, these ODEs could be different from the standard replicator dynamics, (myopic) best response (BR) dynamics, logit dynamics, and fictitious play dynamics. We apply the results to a class of security games in which the attacker and the defender adopt different learning schemes due to differences in their rationality levels and the information they acquire.

AB - Learning algorithms are essential for the applications of game theory in a networking environment. In dynamic and decentralized settings where the traffic, topology and channel states may vary over time and the communication between agents is impractical, it is important to formulate and study games of incomplete information and fully distributed learning algorithms which for each agent requires a minimal amount of information regarding the remaining agents. In this paper, we address this major challenge and introduce heterogeneous learning schemes in which each agent adopts a distinct learning pattern in the context of games with incomplete information. We use stochastic approximation techniques to show that the heterogeneous learning schemes can be studied in terms of their deterministic ordinary differential equation (ODE) counterparts. Depending on the learning rates of the players, these ODEs could be different from the standard replicator dynamics, (myopic) best response (BR) dynamics, logit dynamics, and fictitious play dynamics. We apply the results to a class of security games in which the attacker and the defender adopt different learning schemes due to differences in their rationality levels and the information they acquire.

UR - http://www.scopus.com/inward/record.url?scp=79953139643&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79953139643&partnerID=8YFLogxK

U2 - 10.1109/CDC.2010.5718053

DO - 10.1109/CDC.2010.5718053

M3 - Conference contribution

SN - 9781424477456

SP - 219

EP - 224

BT - 2010 49th IEEE Conference on Decision and Control, CDC 2010

ER -