The Baldwin effect under spatial isolation and autonomous reproduction

The impact of learning on evolution in dynamic environments undergoes recognized stages of the Baldwin Effect although its cause is not clear. To identify it experimentally, we devise spatial constraints and allowed autonomous reproduction for a multiagent game play using Iterated Prisoner's Dilemma. In comparison to Arita and Suzuki's model, we exclude the selective bias of noise, reduce the genotypic and phenotypic correlation and explain how mutant strategies can survive under costs assumptions. Simulation results show that plastic agents are exploited by Defecting, non-plastic agents during game-play. In the majority of the simulations, this Nash Equilibrium Phase occurs, resulting in a Defect-oriented population which eventually wipes itself out. However, with spatial isolation, a small number of cooperative plastic and non-plastic agents survive and proliferate in the simulation after the Defecting agents have died out; the two transitions stages ascribed to the Baldwi...

• Type: Conference
• Year: 2006
• Where: GECCO
• Authors: H. L. Peng, J. C. Tay
• Keywords: Baldwin Effect, GECCO 2006, Non-plastic Agents, Optimization, Prisoner's Dilemma