... evolutionary programming methods can be implemented to study the adaptation of behavior
in a variety of prisoner's dilemmas. ... Evolving Behaviors in the Iterated Prisoner's Dilemma ... Figure
5. The best evolved finite state machine after 25 generations for a given population size. ...

... A lower hound of one move corresponds to the “one-shot” Prisoner's Dilemma. Evolutionary
Computation Volume 3, Number 3 353 Page 6. David B. Fogel All 200 finite-state machines
competed in round-robin pairing (ie, every machine met every other machine once). ...

... known strategies varied between a direct finite state representation for prisoner's dilemma playing
agents, ... The agent representation used in this study is 8-state finite state machines with actions
associated ... Access to state information permits the machine to condition its play on ...

... l. The Prisoner's Dilemma profit; but if two automata achieve the same profit, a metaplayer ... all 26
one and two state Moore machines capable of playing the infinitely repeated Prisoners' Dilemma. ...
B This appendix discusses some formal issues in the theory of finite automata that ...

... following version of the prisoner's dilemma, for example, C D C 2,2 -1,3 D 3,-1 0,0 FIGURE 1. ...
for player i, denoted by Mi, is a four tuple <Qi, qil X1, ti,) where Qi is a finite set, qi ... the machine
is at state qi and the other player chooses s1 e Sj, then the machine's next state is ti(qj, si). ...

... then a finite range of values for x would be selected and the minimum possible ... state spaces can
sometimes be searched quite efficiently, and relatively small state spaces sometimes ... and
engineering problems [74]-[79]; 4) applying genetic algorithms to machine learning rule ...

... this paper have 16 nodes analo- gous to the number of states used in the finite-state machines. ...
but these states are more difficult to access than the states of a finite-state machine as they ... The
action may be a null action, a prisoner's dilemma action (cooperate, defect), or a jump ...

... The finite state machines obtained for both the standard and cellular representations were first
reduced by throwing out all states not ... Evolving behaviors in the iterated prisoners dilemma. ... of
an encounter and the evolution of cooperation in the iterated prisoner's dilemma. ...

... more probe strategies than tit-for-tat, and a proof that the fingerprint of a finite-state machine is
always ... demonstrate that the rate of appearance of several well-known strategies varied between
a direct finite-state rep- resentation for prisoner's dilemma playing agents ...

... The type of finite state machine (Moore or Mealy) affects the sinmlation very little, although there
is a ... Evolutionary dynamics of alturistic behavior in optional and compulsory versions of the iterated
prisoner's dilemma. ... Cooperation in a repeated prisoners' dilemma with ostracism. ...