Visualisation of fuzzy systems: requirements, techniques and framework

Complex fuzzy systems exist in many applications and effective visualisation is required to gain insights into the nature and working of these systems, especially in the implication of imprecision, its propagation and impacts on the quality and reliability of the outcomes. This paper presents a holistic approach towards the design of a visualisation system for fuzzy systems. We firstly analyse the requirements for such a visualisation system by articulating fundamental ontologies that underpin the structure and operations of fuzzy systems. A software framework using a multi-agent approach is then presented with the aim to facilitate the organisation and flow of complex tasks, their inter-relationships and their interactions with users. Finally, we discuss visualisation techniques for fuzzy data and fuzzy rules, and introduce methods to extend and improve some existing techniques. (c) 2004 Elsevier B.V. All rights reserved.