Learning automata: an introduction
Learning automata: an introduction
Robust incentive techniques for peer-to-peer networks
EC '04 Proceedings of the 5th ACM conference on Electronic commerce
A tutorial for competent memetic algorithms: model, taxonomy, and design issues
IEEE Transactions on Evolutionary Computation
Norm diversity and emergence in tag-based cooperation
COIN@AAMAS'10 Proceedings of the 6th international conference on Coordination, organizations, institutions, and norms in agent systems
Emerging cooperation on complex networks
The 10th International Conference on Autonomous Agents and Multiagent Systems - Volume 2
Computers & Mathematics with Applications
Fostering Cooperation through Dynamic Coalition Formation and Partner Switching
ACM Transactions on Autonomous and Adaptive Systems (TAAS)
Hi-index | 0.00 |
This paper presents a framework for describing the spatial distribution and the global frequency of agents who play the spatial prisoner's dilemma with coalition formation. The agent interaction is described by a non-iterated game, where each agent only locally interacts with its neighbours. Every agent may behave as a defector or a cooperator when playing isolated, but they can join or lead coalitions (group of agents) where a leader decides the coalition strategy. Isolated agents' strategies or groups' strategies are public and therefore can be memetically imitated by neighbours. The agent strategy is selected between two possibilities: probabilistic Tit-for-Tat (pTFT) or learning automata (LA). Coalition dynamics are organized around two axes. On the one hand, agents get a percentage of compromise when cooperating with other agents. On the other hand, leaders impose taxes to the other agents belonging to its coalition. These two rules and their related parameters guide the coalition formation and the game evolution. The main contribution of the paper is the framework for memetic analysis of coalition formation in spatial prisoner's dilemma. This work also includes simulation results on a lattice showing that the pTFT memetic approach becomes more effective than an isolated learning policy.