Adaptation of Iterated Prisoner´s Dilemma Strategies by Evolution and Learning

This paper examines the performance and adaptability of evolutionary, learning and memetic strategies to different environment settings in the iterated prisoner´s dilemma (IPD). A memetic adaptation framework is devised for IPD strategies to exploit the complementary features of evolution and learning. In the paradigm, learning serves as a form of directed search to guide evolutionary strategies to attain good strategy traits while evolution helps to minimize disparity in performance between learning strategies. A cognitive double-loop incremental learning scheme (ILS) that encompasses a perception component, probabilistic revision of strategies and a feedback learning mechanism is also proposed and incorporated into evolution. Simulation results verify that the two techniques, when employed together, are able to complement each other´s strengths and compensate each other´s weaknesses, leading to the formation of good strategies that adapt and thrive well in complex, dynamic environments