ALX, AN ACTION LOGIC FOR AGENTS WITH BOUNDED RATIONALITY

We propose a modal action logic that combines ideas from H.A. Simon's bounded rationality, S. Kripke's possible world semantics, G.H. von Wr ight's preference logic, Pratt's dynamic logic, Stalnaker's minimal ch ange and more recent approaches to update semantics. ALX (the xth acti on logic) is sound, complete and decidable, making it the first comple te logic for two-place preference operators. ALX avoids important draw backs of other action logics, especially the counterintuitive necessit ation rule for goals (every theorem must be a goal) and the equally co unterintuitive closure of goals under logical implication.