Reasoning about Actions in a Multiagent Domain Laura Giordano 1 , Alberto Martelli 2 , and Camilla Schwind 3 1 DISTA, Universit`a del Piemonte Orientale “A. Avogadro” laura@mfn.unipmn.it 2 Dipartimento di Informatica - Universit`a di Torino mrt@di.unito.it 3 Facult´ e des Sciences de Luminy, LIM-CNRS schwind@lim.univ-mrs.fr Abstract. In this paper we present a theory for reasoning about ac- tions which is based on the Product Version of Dynamic Linear Time Temporal Logic (denoted DLT L ⊗ ) and allows to describe the behaviour of a network of sequential agents which coordinate their activities by performing common actions together. DLT L ⊗ extends LTL, the propo- sitional linear time temporal logic, by strengthening the until operator by indexing it with the regular programs of dynamic logic. Moreover, it allows the formulas of the logic to be decorated with the names of sequential agents, taken from a finite set. The action theory we propose is an extension of the theory presented in [8], which is based on the logic DLTL, and allows reasoning with incomplete initial states and dealing with postdiction, ramifications as well as with nondeterministic actions. Here we extend this theory to cope with multiple agents synchronizing on common actions. 1 Introduction An approach to reasoning about actions that recently gained renewed attention is the one based on the use of dynamic logic and of temporal logic. The suitability of modal logic for reasoning about actions has been pointed out by several authors in the last years [3,7]. On the one hand, Dynamic Logic [11] adopts essentially the same ontology as McCarthy’s situation calculus, by taking the state of the world as primary, and encoding actions as transformations on states: actions can be represented in a natural way by modalities, and states as sequences of modalities. On the other hand, the adoption of a temporal logic for action theories allows general goals, like achievement and maintenance goals to be specified through temporal modalities. The need of temporally extended goals has been motivated by Bacchus and Kabanza [1] and by Kabanza et al. [14], who proposed an approach to planning based on a linear time temporal logic. The formalization of properties of planning domains as temporal formulas in CTL has also been proposed in [10], where the idea of planning as model checking in a temporal logic has been explored. Other F. Esposito (Ed.): AI*IA 2001, LNAI 2175, pp. 237–248, 2001. c  Springer-Verlag Berlin Heidelberg 2001