5,011 Hits in 4.7 sec

A Self-Organizing Distributed Reinforcement Learning Algorithm to Achieve Fair Bandwidth Allocation for Priority-Based Bus Communication

Tobias Ziermann, Nina Mühleis, Stefan Wildermann, Jürgen Teich
2010 2010 13th IEEE International Symposium on Object/Component/Service-Oriented Real-Time Distributed Computing Workshops  
Based on a game theoretical analysis, a multi-agent reinforcement learning algorithm is proposed that establishes fair bandwidth distribution.  ...  We experimentally evaluated the algorithm with different parameter settings. The algorithm showed to converge to a fair solution in any experiment.  ...  ACKNOWLEDGMENT Supported in part by the German Science Foundation (DFG) under contract TE 163/15-1.  ... 
doi:10.1109/isorcw.2010.18 dblp:conf/isorc/ZiermannMWT10 fatcat:5iscgo7qefeqna72s6syrcbqp4

Analyzing Games with a Variable Number of Players

Madelyn Gatchel
2021 AAAI Conference on Artificial Intelligence  
We hypothesize that the payoffs in a game with x players are similar or related to the same game with x ± 1 players, given a large value of x.  ...  We introduce a novel technique that uses a multi-headed neural network to analyze symmetric games with a variable number of players, where the number of participants falls in a specified range.  ...  This learned deviation payoff function is used in Nash-finding algorithms to find approximate equilibria in simulation-based games with a large number of players, without constructing an explicit payoff  ... 
dblp:conf/aaai/Gatchel21 fatcat:pdt5a6qwx5az7gib3aox5ymy4q

Learning Deviation Payoffs in Simulation-Based Games

Samuel Sokota, Caleb Ho, Bryce Wiedenbeck
We present a novel approach for identifying approximate role-symmetric Nash equilibria in large simulation-based games.  ...  We give a procedure for iteratively refining the learned model with new data produced by sampling in the neighborhood of each candidate Nash equilibrium.  ...  Conclusion We have developed a novel methodology for learning -Nash equilibria in simulation-based games.  ... 
doi:10.1609/aaai.v33i01.33012173 fatcat:afs4ouprvjb2zmhywnzuy4mwou

Neural networks and bounded rationality

Daniel Sgroi, Daniel J. Zizzo
2007 Physica A: Statistical Mechanics and its Applications  
Traditionally the emphasis in neural network research has been on improving their performance as a means of pattern recognition.  ...  Firstly, they select a rule for behavior which appears very similar to that used by laboratory subjects. Secondly, using this rule they perform optimally only approximately 60% of the time.  ...  It is not difficult for the network to find Nash equilibria in specific games, but what is difficult is to learn to employ Nash as a general algorithm.  ... 
doi:10.1016/j.physa.2006.10.026 fatcat:7ndxzlsumzdzfn7ftsgfd3xjqe

Distributive Opportunistic Spectrum Access for Cognitive Radio using Correlated Equilibrium and No-Regret Learning

Zhu Han, Charles Pandana, K. J. Ray Liu
2007 2007 IEEE Wireless Communications and Networking Conference  
To achieve this correlated equilibrium, we construct an adaptive algorithm based on no-regret learning that guarantees convergence.  ...  From the simulation results, the optimal correlated equilibria achieve better fairness and 5%∼15% performance gain, compared to the Nash equilibria.  ...  From the game theory perspective, we propose a distributed protocol based on an adaptive learning algorithm for multiple secondary users using only local information.  ... 
doi:10.1109/wcnc.2007.8 dblp:conf/wcnc/HanPL07 fatcat:dbeejyfqqbbv7airnieydohgea

A Modified Q-Learning Algorithm for Potential Games

Yatao Wang, Lacra Pavel
2014 IFAC Proceedings Volumes  
A Modified Q-Learning Algorithm in Games This thesis presents a modified Q-learning algorithm and provides conditions for convergence to a pure Nash equilibrium in potential games.  ...  In this thesis, we consider a modified Q-learning algorithm based on constant step-sizes, inspired by Joint Strategy Fictitious Play (JSFP).  ...  Chapter 6 Simulations In this chapter, we present simulation results of the modified Q-learning algorithm compared to algorithms discussed in Section 2.2, for an example of a congestion game in a similar  ... 
doi:10.3182/20140824-6-za-1003.02646 fatcat:5eimu66frbhrdkkc7l5tnt3adq

Computational and Data Requirements for Learning Generic Properties of Simulation-Based Games [article]

Cyrus Cousins, Bhaskar Mishra, Enrique Areyan Viqueira, Amy Greenwald
2022 arXiv   pre-print
Empirical game-theoretic analysis (EGTA) is primarily focused on learning the equilibria of simulation-based games.  ...  We conclude with experiments that uncover some of the remaining difficulties with learning properties of simulation-based games, in spite of recent advances in statistical EGTA methodology, including those  ...  still learning 2ε-approximate equilibria of the simulation-based game.  ... 
arXiv:2208.06400v1 fatcat:ove64thzjfegbia3jvfrcluh54

Solving Graph-based Public Good Games with Tree Search and Imitation Learning [article]

Victor-Alexandru Darvariu, Stephen Hailes, Mirco Musolesi
2021 arXiv   pre-print
Existing algorithms for this known NP-complete problem find solutions that are sub-optimal and cannot optimize for criteria other than social welfare.  ...  In particular, we define a Markov Decision Process which incrementally generates an mIS, and adopt a planning method to search for equilibria, outperforming existing methods.  ...  For graph-based best-shot PGGs, it has been shown that each equilibrium ( ) Figure 1 : Schematic of our approach for finding desirable equilibria in the graph-based best-shot game.  ... 
arXiv:2106.06762v2 fatcat:zgymmvtgb5fu5j6ykimxgzydeu

An algorithmic game theory study of wholesale electricity markets based on central auction

Sofia Ceppi, Nicola Gatti, Francesco Amigoni, Maria Gini, Wolfgang Ketter
2010 Integrated Computer-Aided Engineering  
The second approach, based on multi-agent simulations, assumes that generators implement simple learning algorithms.  ...  In this paper, we provide an algorithmic game theory study that improves the state of the art related to both the two previous approaches.  ...  We provided an algorithmic game theory study that improves the state of the art related to both the microeconomic works and the multi-agent based simulation ones.  ... 
doi:10.3233/ica-2010-0346 fatcat:kho5eckt6jbklbh7qgmqbtplye

Two Approaches to Building Collaborative, Task-Oriented Dialog Agents through Self-Play [article]

Arkady Arkhangorodsky, Scot Fang, Victoria Knight, Ajay Nagesh, Maria Ryskina, Kevin Knight
2021 arXiv   pre-print
We give empirical results for both reinforcement learning and game-theoretic equilibrium finding.  ...  However, human/human corpora are frequently too small for supervised training to be effective.  ...  Likewise, the user cannot improve Figure 4 : Equilibria computed for the game tree in Figure 3 , giving strategies for User and Agent.  ... 
arXiv:2109.09597v1 fatcat:vjyxsflsb5givhv6luefjpyoj4

Analyzing Oligopolistic Electricity Market Using Coevolutionary Computation

H. Chen, K.P. Wong, D.H.M. Nguyen, C.Y. Chung
2006 IEEE Transactions on Power Systems  
This paper presents a new unified framework of electricity market analysis based on coevolutionary computation (CCEM) for both the one-shot and the repeated games of oligopolistic electricity markets.  ...  The standard Cournot model and the new Pareto improvement model are used. The linear and constant elasticity demand functions are considered.  ...  CCEM for Pareto Improvement Model We use the algorithm described in Section III-C to perform the simulation.  ... 
doi:10.1109/tpwrs.2005.862005 fatcat:otxpcrpfvjbajhwsg6q3invhuq

Agent-based Simulation for Research in Economics [chapter]

Clemens van Dinther
2008 Handbook on Information Technology in Finance  
Nevertheless, the methodology of economic simulations in general and agent-based simulations in particular is still in question.  ...  Agent-based simulations have produced a variety of interesting contributions to economic research in the past years.  ...  [28] have introduced an algorithm for multi-agent learning in two player games.  ... 
doi:10.1007/978-3-540-49487-4_18 fatcat:w23rusoswzeulcc54ijgjcm57m

Simultaneous Adversarial Multi-Robot Learning

Michael H. Bowling, Manuela M. Veloso
2003 International Joint Conference on Artificial Intelligence  
It combines gradient-based policy learning techniques with the WoLF ("Win or Learn Fast") variable learning rate. We apply this algorithm to an adversarial multirobot task with simultaneous learning.  ...  In this paper we introduce GraWoLF, a general-purpose, scalable, multiagent learning algorithm.  ...  The authors also thank Brett Browning and James Bruce for the development of the CMDragons'02 robots used in this work.  ... 
dblp:conf/ijcai/BowlingV03 fatcat:tnzvbzb7lvdvvbmsvbdhpgl7vy

Artificial Intelligence and Auction Design [article]

Martino Banchio, Andrzej Skrzypacz
2022 arXiv   pre-print
Motivated by online advertising auctions, we study auction design in repeated auctions played by simple Artificial Intelligence algorithms (Q-learning).  ...  We show that the difference is driven by the incentive in first-price auctions to outbid opponents by just one bid increment.  ...  Our theory of repeated games suggests that improved monitoring and speed of reaction make tacit-collusive equilibria more stable.  ... 
arXiv:2202.05947v1 fatcat:pdnxp76j6jby7gxbeughh6fc3u

Spatial spectrum access game

Xu Chen, Jianwei Huang
2012 Proceedings of the thirteenth ACM international symposium on Mobile Ad Hoc Networking and Computing - MobiHoc '12  
Numerical results demonstrate that the distributed learning algorithm achieves up to 100% performance improvement over a random access algorithm.  ...  We show that the distributed learning algorithm can converge to an approximate mixed-strategy Nash equilibrium for any spatial spectrum access games.  ...  Distributed Learning Algorithm Based on the expected throughput estimation, we now propose the distributed learning algorithm for spatial spectrum access games.  ... 
doi:10.1145/2248371.2248401 dblp:conf/mobihoc/ChenH12 fatcat:nnnlswitcfd6tfix5mtd3smt5a
« Previous Showing results 1 — 15 out of 5,011 results