Fuzzy Reasoning with Ontology: Towards Truly Autonomic Semantic Web Service Consumption

To ensure autonomous consumption of services by software agents, Web Services have to be represented in a machine-understandable form. With this infrastructure in place, agents acting on behalf of their human users can automatically locate, discover, compose, and execute required services. Such scenario is based on the recently introduced concept of Semantic Web: an agent should know personal preferences of its user, and use them to find and engage services providing the best match to these preferences. User preferences can be represented using ontologies. Conventional ontologies, however, do not provide means for representing concepts that are vague or approximate, as typical for humans. Similarly, conventional matching mechanisms may not provide the best match as perceived by users. In this paper, ontology is extended by concepts of fuzziness and matching mechanism by methods of approximate reasoning. Such approach aims at providing capability to mimic human performance in multi-criteria decisionmaking, as illustrated in a simple application.