Matchmaking arises when supply and demand meet in an electronic marketplace, or when agents search for a web service to perform some task, or even when recruiting agencies match curricula and job profiles. In such open environments, the objective of a matchmaking process is to discover best available offers to a given request. We address the problem of matchmaking from a knowledge representation perspective, with a formalization based on Description Logics. We devise Concept Abduction and Concept Contraction as non-monotonic inferences in Description Logics suitable for modeling matchmaking in a logical framework, and prove some related complexity results. We also present reasonable algorithms for semantic matchmaking based on the devised inferences, and prove that they obey to some commonsense properties. Finally, we report on the implementation of the proposed matchmaking framework, which has been used both as a mediator in e-marketplaces and for semantic web services discovery.
The promise of the Semantic Web initiative is to revolutionize the way information is coded, stored, and searched on the Internet (Berners-Lee, Hendler, & Lassila, 2001). The basic idea is to structure information with the aid of markup languages, based on the XML language, such as RDF and RDFS 1 , and OWL 2 . These languages have been conceived for the representation of machine-understandable, and unambiguous, description of web content through the creation of domain ontologies, and aim at increasing openness and interoperability in the web environment.
Widespread availability of resources and services enables-among other advantagesthe interaction with a number of potential counterparts. The bottleneck is that it is difficult finding matches, possibly the best ones, between parties.
The need for a matchmaking process arises when supply and demand have to meet in a marketplace, or when web services able to perform some task have to be discovered, but also when recruiting agencies match curricula and job profiles or a dating agency has to propose partners to a customer of the agency. Requests and offers may hence be generic demands and supplies, web services, information, tangible or intangible goods, and a matchmaking process should find for any request an appropriate response. In this paper we concentrate on automated matchmaking, basically oriented to electronic marketplaces and service discovery, although principles and algorithms are definitely general enough to cover also other scenarios. We assume, as it is reasonable, that both requests and offers are endowed of some kind of description. Based on these descriptions the target of the matching process is finding, for a given request, best matches available in the offers set, and also, given an offer, determine best matching requests in a peer-to-peer fashion. We may hence think of an electronic mediator as the actor who actively tries to carry out the matchmaking process. Obviously descriptions might be provided using unstructured text, and in this case such an automated mediator should revert to adopting either basic string matching techniques or more sophisticated Information Retrieval techniques.
The Semantic Web paradigm calls for descriptions that should be provided in a structured form based on ontologies, and we will assume in what follows that requests and offers are given with reference to a common ontology. It should be noticed that even when requests and offers are described in heterogeneous languages, or using different ontologies modelling the same domain, schema/data integration techniques may be employed to make them comparable, as proposed e.g., by Madhavan, Bernstein, and Rahm (2001), and Shvaiko and Euzenat (2005); but once they are reformulated in a comparable way, one is still left with the basic matchmaking problems: given a request, are there compatible offers? If there are several compatible offers, which, and why, are the most promising ones?
Matchmaking has been widely studied and several proposals have been made in the past; we report on them in Section 2. Recently, there has been a growing effort aimed at the formalization with Description Logics (DLs) (Baader, Calvanese, Mc Guinness, Nardi, & Patel-Schneider, 2003) of the matchmaking process (e.g., Di Sciascio, Donini, Mongiello, & Piscitelli, 2001;Trastour, Bartolini, & Priest, 2002;Sycara, Widoff, Klusch, & Lu, 2002;Di Noia, Di Sciascio, Donini, & Mongiello, 2003b;Li & Horrocks, 2003;Di Noia, Di Sciascio, Donini, & Mongiello, 2003c, 2003a, among others). DLs, in fact, allow to model structured descriptions of requests and offers as concepts, usually sharing a common ontology. Furthermore DLs allow for an open-world assumption. Incomplete information is admitted, and absence of information can be distinguished from negative information. We provide a little insight on DLs in Section 3.
Usually, DL-based approaches exploit standard reasoning services of a DL systemsubsumption and (un)satisfiability-to match potential partners in an electronic transaction. In brief, if a supply is described by a concept Sup and a demand by a concept Dem, unsatisfiability of the conjunction of Sup and Dem (noted as Sup ⊓ Dem) identifies the incompatible proposals, satisfiability identifies potential partners-that still have to agree on underspecified constraints-and subsumption between Sup and Dem (noted as Sup ⊑ Dem) means that requirements on Dem are completely fulfilled by Sup.
Classification into compatible and incompatible matches can be useless in the presence of several compatible supplies; some way to rank most promising ones has to be identified; also some explanation on motivation of such a rank could be appreciated. On the other hand, when there is lack of compatible matches one may accept to turn to incompatible matches that could still be interesting, by revising some of the original requirements presented in the request, as far as one could easily identify them.
In other wo
This content is AI-processed based on open access ArXiv data.
