A Theory of Bounded Inductive Rationality

The dominant theories of rational choice assume logical omniscience. That is, they that when facing a decision problem, an agent can perform all relevant computations and determine the truth value logical/mathematical claims. This assumption is unrealistic when, for example, we offer bets on remote digits pi or faces computationally intractable planning problem. Furthermore, omniscience creates contradictions in cases where environment contain descriptions itself. Importantly, strategic interactions as studied game theory are problems which predicted by its (the other players). In this paper, develop making does not We consider agents who repeatedly face (including ones like betting games against agents). main contribution paper to provide sensible rationality such agents. Roughly, require boundedly inductive tests each efficiently computable hypothesis infinitely often follows those hypotheses keep their promises high rewards. then prove sense have desirable properties. For learn random pseudo-random lotteries at expected reward. Finally, between different folk theorem what strategies bounded converge to.