Applying Fogel and Burgin´s `Competitive goal-seeking through evolutionary programming´ to coordination, trust, and bargaining games

For original paper, see L.J. Fogel and G.H. Burgin, Final Report, Contract AF 19(628), Air Force Cambridge Research Labs (1969). Experiments are conducted in which a population of finite state machines learn to play three different games termed “coordination”, “trust” and “bargaining”. These games were offered in the early literature of evolutionary computation and can serve as possible benchmarks for study. In particular, attention is focused on the emergent behavior that arises when a population of finite state machines start with random connectivity and must learn appropriate moves in the light of the composition of the other members of the population. The results found by extending the early efforts in the trust game, also known as the prisoner´s dilemma, match well with seminal results offered by R. Axelrod (1987)