We present a framework for reasoning about processes (complex actions) that
are constituted by several concurrent activities performed by various inte
racting agents, The framework is based on two distinct formalisms: a repres
entation formalism, which is a CCS-like process algebra associated with an
explicit global store; and a reasoning formalism, which is an extension of
modal mu-calculus, a powerful logic of programs that subsumes dynamic logic
s such as PDL and Delta PDL, and branching temporal logics such as CTL and
CTL*. The reasoning service of interest in this setting is model checking i
n contrast to logical implication. This framework, although directly applic
able only when complete information on the system behavior is available, ha
s several interesting features for reasoning about actions in Artificial In
telligence. Indeed, it inherits formal and practical tools from the area of
Concurrency in Computer Science, to deal with complex actions, treating su
itably aspects like nonterminating executions, parallelism communications,
and interruptions. (C) 1999 Published by Elsevier Science B.V. All rights r
eserved.