Abstract:
With the increasing popularity of Web services, number of available services on public and business domains grows rapidly in the recent years. This growth in the number of services makes service discovery more important and challenging. The fundamental element of service discovery is service matchmaking. Service matchmaking is the process of retrieving suitable services given by service providers to satisfy service request of service consumers. The current standard service discovery mechanism provides only primitive service matchmaking methods, which are not su cient to ful ll the requirements of todays consumers. Recent research to develop better service matchmaking methods is based on the use of semantic models of input-output interfaces of services and their use in service matchmaking. However these methods su er from low precision due to lack of use of internal process knowledge of services in matchmaking. In this thesis we propose two novel service matchmaking methods to achieve better precision than the state of the art service matchmaking methods. In order to achieve this goal, we claim that extensive use of semantically augmented internal process knowledge of service is necessary. Hence, in our proposed methods we use internal process models, which we markup with semantic concepts, as the core information source for service matchmaking. Our rst matchmaking method uses nite state machines for service modeling and several structural and a semantic similarity metric for matchmaking. Our second method uses temporal logic for modeling and model checking techniques for matchmaking. We propose a generic service matchmaking framework to realize our proposed approaches and conduct case studies to evaluate strong and weak points of our proposed matchmaking methods.
