Abstract not found

When giving an nalysis of knowledge in multiagent systems, one needs a framework in which higher-order information and its dynamics can both be represented. A recent tradition stoxting in origina work by Plaza treats all of knowledge, higher-order knowledge, and its dynamics on the sae foot. Our work is in that tradition. It also fits in approaches that not only dynaize the epistemics, but also epistemize the dynamics: the ac- tions that (groups of) agents perform oxe epistemic actions. Different agents may have different information about which action is taking place, including higher-order information. We demonstrate that such information changes require subtle descriptions. Our contribution is to provide a complete axiomatization for n action language of vn Ditmoxsch, where an action is interpreted as a relation between epistemic states (pointed models) and sets of epistemic states. The applicability of the framework is found in every context where multiagent strategic decision making is at stake, and aready demonstrated in gae-like scenoxios such as Cluedo and coxd games.

The notion of preference occurs across many areas, including the philosophy of action, decision theory, optimality theory, and game theory. In these settings, individual preferences between worlds or actions can be used to predict behavior by rational agents. In a more abstract sense, the notion of preference also

In this paper, we introduce a notion of "epistemic action" to describe changes in the information states of the players in a game. For this, we use ideas that we have developed in our previous papers [BMS], [BMS2] and [B], enriching them to cover, not just purely epistemic actions, but also fact-changing actions ("real moves", e.g. choosing a card, exchanging cards etc.) and nondeterministic actions and strategies (conditional actions having knowledge tests as conditions). We consider natural operations with epistemic actions and we use them to describe significant aspects of the interaction between beliefs and actions in a game. For this, we use a logic that combines in a specific way a multi-agent epistemic logic with a dynamic logic of "epistemic actions". We give (without proof) a complete and decidable proof system for this logic. As an application, we analyze a specific example of a dialogue game (a version of the Muddy Children Puzzle, in which some of the children can ...

ABSTRACT. We show how belief revision can be treated systematically in the format of dynamicepistemic logic, when operators of conditional belief are added. The core engine consists of definable update rules for changing plausibility relations between worlds, which have been proposed independently in the dynamic-epistemic literature on preference change. Our analysis yields two new types of modal result. First, we obtain complete logics for concrete mechanisms postulates for belief revision can be analyzed by standard modal frame correspondences for model-changing operations.

Abstract not found

1.1 General issues........................... 5

This paper studies coalgebras from the perspective of finite observations. We introduce the notion of finite step equivalence and a corresponding category with finite step equivalence-preserving morphisms. This category always has a final object, which generalises the canonical model construction from Kripke models to coalgebras. We then turn to logics whose formulae are invariant under finite step equivalence, which we call logics of rank . For these logics, we use topological methods and give a characterisation of compact logics and definable classes of models.

abstract. We survey a number of decidablity and undecidablity results concerning epistemic temporal logic. The goal is to provide a general picture which will facilitate the ‘sharing of ideas ’ from a number of different areas concerned with modeling agents in interactive social situations. 1

Dynamic logic is a logic to reason about the dynamics of (natural or artificial) systems in general, ranging from the effects of actions of human agents to the behaviour of artificial agents and software systems. Therefore it is to be expected that in AI it can be fruitfully employed both to represent knowledge about the dynamics of the domain at hand as well as to describe / specify (the dynamic behaviour of) AI systems themselves. A typical example of the former is the description of the effects of actions (of humans, for example) in the commonsense world, while the specification of a particular reasoning system would be of the latter type. In this paper a number of examples are given to illustrate the usefulness (and wide scope!) of dynamic logic for AI. 1 Introduction Originally, dynamic logic has been proposed in computer science as a logic for reasoning about programs in order to verify their correctness. Later it was realised that dynamic logic could also be fruitful...