Approximate reasoning and Semantic Web Services

The initiative of representing Web Services in a machine-understandable way creates a new era for software agent interoperability. A recent introduction of a concept of the Semantic Web makes it possible to automatically locate, discover, composite, and execute the services. A user agent works on behalf of its owner, and knows user´s personal preferences. As there might be many service providers on the web, finding the best one which matches user´s needs is critical for a suitable performance of the agent. In the paper, fuzziness and approximate reasoning methodology are applied within the Semantic Web environment. The proposed approach aims at providing capability to mimic human behavior in the case of a multi-criteria decision making process. Ontology with fuzziness is used to represent human needs and preferences. This ontology contains also information about different acceptance levels which user may have depending on responses obtained from different service providers. A prototype Semantic Web Service representing hotel reservation is built using the approach proposed.