Contact Details

Dr. Barry McCollum

Dr. Paul McMullan

Modelling Human Reasoning within the Assessment Process

Fuzzy logic and fuzzy methods were introduced by Zadeh in the 1960’s and 1970’s and have had great success in areas of control engineering. Indeed, ‘fuzzy logic control’ has been used in consumer electronic devices such as washing machines, car automatic gearboxes and auto-focus cameras. Given that one of the original motivations behind the introduction of fuzzy methods was to model human reasoning, it is somewhat ironic that less progress has been made in this area. Recent developments in extensions to the basic framework of standard (type-1) fuzzy logic have begun to have an impact in this area by providing more flexible computational tools for modelling fuzzy sets which themselves contain uncertainty.

In practical examination timetabling there is a need to address the real requirements of users when assessing the quality of automatically generated timetable solutions. Unfortunately, there has so far been little progress in meeting this need. The aim of this PhD project is to bring together the latest advances in extensions to type-1 fuzzy logic with the requirements of real timetabling officers in order to generate novel methods for extracting and modelling knowledge, carrying out fuzzy-based reasoning and producing interpretable results in order to provide a decision support system to guide users in assessing the relative qualities of alternative timetable solutions. Novel fuzzy computational techniques in how to extract knowledge and represent this in non-standard fuzzy sets will be required.