This report describes a Java API, created at the National Research Council of Canada (NRC), that provides the capability for handling fuzzy concepts and reasoning. We call this the NRC FuzzyJ Toolkit. It is hoped that this proves useful for exploring the ideas of fuzzy logic and fuzzy reasoning in a Java setting. Much of the work is based on earlier experience building the FuzzyCLIPS extension to the CLIPS Expert System Shell. A set of classes is also included to allow an integration of the FuzzyJ Toolkit and Jess, the Expert System Shell, from Sandia National Labs.
This documentation does not attempt to show how one designs and builds fuzzy reasoning systems. It only attempts to show the fundamentals of using the toolkit itself, describing the major classes and providing some simple examples of the use of these classes. It assumes at least some familiarity of fuzzy logic and its terminology as well as Java (and Jess where required).
The document should be read in conjunction with the API documentation for complete understanding. Browsing the examples provided will also help to see some techniques for building fuzzy systems. There is great flexibility with the toolkit and many unexpected combinations will likely be discovered by astute users of the system. Feel free to share them with us and feel free to offer suggestions and report problems.