Genetic and evolutionary algorithms come of age
Communications of the ACM
Prisoner's Dilemma
Modelling agent strategies in simulated market using iterated prisoner's dilemma
ICCSA'07 Proceedings of the 2007 international conference on Computational science and its applications - Volume Part III
Hi-index | 0.00 |
This paper investigates the use of Genetic Algorithms (GA) to evolve cooperative agents in a competitive market environment using Iterated Prisoner's Dilemma (IPD). Our study seeks to follow Axelrod's research of computer simulations of the IPD game which is generally regarded as a benchmark for the studies on evolution of cooperation. However, we are of the opinion that his work was a little restrictive and lack of a genuine real-world component. In this paper, we report on a simulation study that attempts to bridge the gap by applying GA to a market model. We examine how well GA could perform against the IPD strategies, and explore the strategic interactions among the agents that represent firms in a coevolving population. We also report on the influence of the genetic operators on the performance of GA. Our experimental results show that cooperation can be evolved in such non-cooperative environment using GA. We conclude that with proper tuning of parameters, GA could be extremely useful for optimizing the outcome of an economy market.