Verification of Multi-agent Systems with Imperfect Information and Public Actions

Abstract : We analyse the verication problem for synchronous, perfectrecall multi-agent systems with imperfect informationagainst a specication language that includes strategic andepistemic operators. While the verication problem is undecidable,we show that if the agents' actions are public, thenverication is 2exptime-complete. To illustrate the formalframework we consider two epistemic and strategic puzzleswith imperfect information and public actions: the muddychildren puzzle and the classic game of battleships.