Tuning Costraint-Based Communication in Distributed Problem Solving
Francesca Arcelli Fontana, Uwe M. Borghoff, Ferrante Raffaele Formato, Remo Pareschi · 1995
@xerox.frAbstract Distributed Problem Solving (DPS) decomposes problems into subproblems to besolved by interacting, cooperative software agents. Thus, DPS is suitable for modeling, in thecontext of parallel and distributedarchitectures, the solvingof problems characterized by manyinter-dependencies among subproblems. Concurrent Constraint Programming (CCP) providesa powerful execution framework for DPS, where constraints can declaratively implement bothlocal problem solving as well as exchange of information, and hence DPS, among agents. Tooptimize DPS, the protocol for constraint communication must be tuned to the specific kindof DPS problem and the characteristics of the underlying system architecture. In this paper,we provide a formal framework for modeling different options and we show how it applies toconcrete, generalizable examples.Key words: constraintpropagation,distributedartificial intelligence,distributedproblemsolv-ing, constraint-based knowledge brokers, cooperative agents, protocols.