Proceedings of the 2004 IEEE/WIC/ACM International Conference on Web Intelligence

Quantitative Agent Service Matching

, , and

The ultimate goal of service matching is to find the service provider(s) that can perform tasks described in a given request with the highest overall satisfaction. However, service description matching solves only part of the problem. Agents that match a given service request description may vary greatly in the level and/or quality of services that they can perform, and an agent may have strong and weak areas in its advertised service space. In this work, we take a quantitative approach in which the broker agent (matchmaker) treats performance ratings as an integral part of an agent’s capability model, captures an agent’s strengths and weaknesses through interaction with the agents, and refines the agent’s capability model with the information gathered. The broker agent also dynamically builds a service distribution model that provides vital information for determining the degree of match in non-exact cases. We designed and implemented an experimental system using OWL-S as the upper service ontology, and the results show a significant advantage over other major broker levels.


  • 229787 bytes

ai, computer industry, computer science, marketing and sales, ontology, owl, performance evaluation, portable computers, rdf, semantic web, software agents, statistics

InProceedings

IEEE

DOI: 10.1109/WI.2004.10169

Downloads: 388 downloads

Google Scholar Citations: 4 citations

UMBC ebiquity