Intelligent Production Machines and Systems D. T. Pham, E. E. Eldukhri and A. J. Soroka (eds) © 2005 Elsevier Ltd. All rights reserved Hypermedia-based adaptive expert system for advanced performance support D. T. Pham a , A. M. Huneiti b 3 Manufacturing Engineering Centre, Cardiff University,Wales, UK, E-mail: PhamDT@cardiff.ac.uk h Information Technology Department, CIS division, Jordan University, Amman-Jordan, E-mail: A.huneiti@ju.edu.jo Abstract This paper discusses the provision of advanced performance support through the integration of techniques used in developing knowledge-based systems, namely diagnostic expert systems and adaptive hypermedia systems. This integration is implemented by employing hypermedia which allows supporting content to be synchronized with the diagnostic expert system inference process. The paper proposes an integrated fault data model for diagnostic expert systems. It also introduces a technique for the adaptive retrieval of hypermedia- based technical information using data semantics. A special organisation of displays in the user interface of the adaptive diagnostic expert system allows operators, while applying the expert system for fault diagnosis, to request detailed information about a certain diagnosis procedure, and then return to the system to continue from where they left off. The techniques proposed in this paper are demonstrated through a prototype adaptive expert system for locating and correcting braking system faults in a forklift truck. 1. Introduction Supporting the performance of workers in modern high technology job environments has become an increasingly complex, time consuming and costly task that requires advanced methods. Simple tasks of low information volume can be efficiently facilitated using traditional performance support methods such as paper documentation, electronic databases, lectures, instructor-led courses, and job aids. However, many problems are associated with these traditional methods especially when used to assist the operation of complex products or systems. These concern their portability, complexity, accuracy, reliability, and information maintainability [1]. There are also other problems associated with integrating and retrieving the information, the static structure of the presented material, the restricted support given to users, and the limitations of presentation methods [2]. Reliable Performance Support Systems (PSS) can enhance productivity and reduce training, time and costs and operating errors. Moreover, they can increase quality, task completion rate and worker autonomy [3, 4]. However, according to Fischer and Horn [5], without tools that are primarily PSS tools and with no clear methodology for building them or measuring their performance, PSSs will be limited to being just "an approach". This paper discusses the provision of advanced performance support through the utilisation and integration of technologies used in developing knowledge-based systems, namely diagnostic expert systems and adaptive hypermedia systems. Section 2 reviews advanced performance support systems. Section 3 introduces an integrated fault data model which enables the automatic generation of the knowledge base of a diagnostic expert system. Section 4 presents a methodology for the adaptive retrieval of technical information using data semantics. This methodology is supported by a stereotype user model and implemented using conditional semantic rules. Section 5 demonstrates 249