Inscriere cercetatori

Site nou !

Daca nu va puteti recupera parola (sau aveti alte probleme), scrieti-ne la pagina de contact. Situl vechi se gaseste la adresa old.ad-astra.ro

Facebook

The convergence of the Iterated Prisoner’s Dilemma game

Domenii publicaţii > Ştiinţe informatice + Tipuri publicaţii > Articol în revistã ştiinţificã

Autori: Leslie A. Goldberg, Catherine Greenhill, Martin Dyer, Gabriel Istrate and Mark Jerrum

Editorial: Combinatorics, Probability and Computing, 11, p.135-157, 2002.

Rezumat:

We investigate a simple dynamics on graphs, motivated by:
1. The Pavlov rule in Prisoner’s Dilemma
2. A special case of co-learning, a concept in distributed Artificial Intelligence introduced by Shoham and Tennenholtz , and further investigated by Kittock.

We prove two results:
1. On cycles absorbtion takes O(n log(n)) steps.
2. In contrast, the number of steps is exponential on the
complete graph.

Cuvinte cheie: Prisoner's Dilemma, Markov chains, mixing time

URL: http://dx.doi.org/doi:10.1017/S096354830100503X