Growing artificial societies: social science from the bottom up
Growing artificial societies: social science from the bottom up
Adaptive Learning by Genetic Algorithms: Analytical Results and Applications to Economic Models
Adaptive Learning by Genetic Algorithms: Analytical Results and Applications to Economic Models
Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control and Artificial Intelligence
Genetic Algorithms in Search, Optimization and Machine Learning
Genetic Algorithms in Search, Optimization and Machine Learning
On Evolving Robust Strategies for Iterated Prisoner's Dilemma
AI '93/AI '94 Selected papers from the AI'93 and AI'94 Workshops on Evolutionary Computation, Process in Evolutionary Computation
Two essays on the economics of imperfect information
Two essays on the economics of imperfect information
Simula Begin
Dynamic Plots in Virtual Negotiations
Computational & Mathematical Organization Theory
Adaptive Trust and Co-operation: An Agent-Based Simulation Approach
Proceedings of the workshop on Deception, Fraud, and Trust in Agent Societies held during the Autonomous Agents Conference: Trust in Cyber-societies, Integrating the Human and Artificial Perspectives
Experimentation and Learning in Repeated Cooperation
Computational & Mathematical Organization Theory
Robust Evolutionary Algorithm Design for Socio-economic Simulation
Computational Economics
Adaptive learning in complex trade networks
SEAL'06 Proceedings of the 6th international conference on Simulated Evolution And Learning
| Hi-index | 0.00 |
A Prisoner‘s dilemma that is repeated indefinitely has manyequilibria; the problem of selecting among these is often approachedusing evolutionary models. The background of this paper is a numberof earlier studies in which a specific type of evolutionary model, agenetic algorithm (GA), was used to investigate which behaviorsurvives under selective pressure. However, that normative instrumentsearches for equilibria that may never be attainable. Furthermore, itaims for optimization and, accordingly, says what people should do to be successful in repeated prisoner‘s dilemma (RPD) typesituations. In the current paper, I employ simulation to findout what people would do, whether this makes them successful ornot. Using a replication of Miller‘s (1988) GA study for comparison,a model is simulated in which the population is spatially distributedacross a torus. The agents only interact with their neighbors andlocally adapt their strategy to what they perceive to be successfulbehavior among those neighbors. Although centralized GA-evolution maylead to somewhat better performance, this goes at the cost of a largeincrease in required computations while a population withdecentralized interactions and co-adaptation is almost as successfuland, additionally, endogenously learns a more efficient scheme foradaptation. Finally, when the agents‘ perceptive capabilities arelimited even further, so that they can only perceive how theirneighbors are doing against themselves, rather than against all thoseneighbors‘ opponents—which essentially removes reputation as asource of information—cooperation breaks down.