Resource matchmaking based on negotiable and strict constraints in the semantic web

As more and more services and deployed software agents populate theWeb, the need for effective and automated search and selection of relevant resources becomes of paramount importance both for humans and automated agents as well. Semantic-based search and selection of resources, described according to a structured description, is usually called matchmaking. In this paper we extend the features of semantic matchmaking, by exploiting recently proposed nonstandard inference services in Description Logics. Our approach allows to manage negotiable and strict constraints of a requested resource (or an offered one) while performing a matchmaking process, even if both the request and the offer are incompatible –some part of one description is in conflict with the other– and some constraints in one description are not specified in the other one. An algorithm is presented to compute both which part of the request should be retracted and which part of the description has to be refined in order to make them completely satisfiable with each other. The advantage of the approach is two-fold: it opens negotiation spaces between service providing and service requesting agents, and provides explanation on which constraints have to be relaxed or are under-specified.