Abstract | ||
---|---|---|
In this paper, we study the conditions under which mutual cooperation among one-step memory evolutionary agents instochastic iterated prisoner's dilemma game platform can be sustained. The agents evolve along the gradient of the payoff field. A metric is introduced to quantify players' ability to sustain cooperation with other players. Numerical experiment indicates each allele's role in sustaining the cooperative relationship by hiding its weaknesses to the opponent. The results were analyzed mathematically via transforming the problem into Markov chain state value problem to obtain partial derivatives of payoff functions. Finally, possible applications of methodology and results in this paper to multi-agent systems control are discussed.
|
Year | DOI | Venue |
---|---|---|
2015 | 10.1145/2695664.2696031 | SAC 2015: Symposium on Applied Computing
Salamanca
Spain
April, 2015 |
Keywords | Field | DocType |
Prisoner's Dilemma, Reinforcement learning, Multi-agent system, Cooperation, Artificial Life | Mathematical economics,Computer science,Markov chain,Prisoner's dilemma,Multi-agent system,Artificial intelligence,Dilemma,Traveler's dilemma,Superrationality,Reinforcement learning,Stochastic game | Conference |
ISBN | Citations | PageRank |
978-1-4503-3196-8 | 0 | 0.34 |
References | Authors | |
0 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Kim Minsam | 1 | 0 | 0.34 |
Kwok Yip Szeto | 2 | 64 | 21.47 |