Fuzzy matchmaking in e-marketplaces of peer entities using datalog

We present an approach to matchmaking in electronic marketplaces of peer entities, which mixes in a formal and principled way Datalog, fuzzy sets and utility theory, in order to determine the most promising matches between prospective counterparts. The use of Datalog ensures the scalability of our approach to large marketplaces, while fuzzy logic provides a neat connection with logical specifications and allows to model soft constraints and how well they could be satisfied by an agreement. Noteworthy is that our approach takes into account in the peer-to-peer matchmaking also preferences of each counterpart and their utilities. This allows to rule out of the match list those counteroffers that, although seemingly appealing for the buyer, would probably lead to failure due to contrasting preferences of the seller, and paves the way to the actual negotiation stage.