Connguration as Composite Constraint Satisfaction

Selecting and arranging parts is the core of a conngu-ration task. The validity of a connguration is deened in terms of constraints. Highly declarative, domain independent and simple to use, the constraint satisfaction problem (CSP) paradigm ooers an adequate framework for this task. However, the basic paradigm is not powerful enough to capture or to take advantage of essential aspects of connguration, such as the unknown a priori number of constituent parts of a system or the inherent internal structure of these parts. Although notable eeort has been spent on extending the basic paradigm to accommodate these issues, we still lack a comprehensive formalism for connguration. This paper presents the main ideas behind a general constraint-based model of connguration tasks represented as a new class of nonstandard constraint satisfaction problems, called composite CSP. Composite CSP uniies several CSP extensions, providing a more comprehensive and eecient basis for formulating and solving connguration problems.