Authors: G. Shorbagy; M. Zaki; M. Maree; A. Rafea
Addresses: College of Computer and Information Sciences, King Saud University, Riyadh, KSA ' Department of Systems and Computers, Faculty of Engineering, Azhar University, Cairo, Egypt ' Department of Systems and Computers, Faculty of Engineering, Azhar University, Cairo, Egypt ' Department of Computer Sciences and Engineering, American University in Cairo, Cairo, Egypt
Abstract: On the web, finding a service which is suited to user requirements is essential. In this paper, a multi-agent personalisation of the SWS system, MPSWS, is introduced. In MPSWS, both user information and SWS are interacting to identify user needs. It exploits agent-based systems to support the interaction between users and semantic web services in dynamic configurations. The work achieves the following contributions: 1) proposing a novel system architecture that realises personalisation in the semantic web services; 2) applying an AQ machine learning algorithm to proactively identifying user intentions from past events; 3) providing a BDI-based problem solver that exploits the user's context to determine his intentions; 4) building up OWLS queries for each identified user intention, which are processed by SWS matchmaker, to identifying OWLS services that can semantically match user needs. MPSWS is experimentally tested and its performance at different operational conditions is evaluated.
Keywords: semantic web; web services; SWS; agent technology; personalisation; ontology; context awareness; BDI agents; belief desire intention; AQ machine learning; multi-agent systems; MAS; agent-based systems; user information; user intentions; OWLS queries.
International Journal of Web Science, 2014 Vol.2 No.4, pp.258 - 286
Received: 24 Jan 2014
Accepted: 09 Jan 2015
Published online: 15 Jul 2015 *