VAST Improvements to Diagrammatic Scheduling Using Representational Epistemic Interface Design

Abstract

REpresentational EPistemic Interface Design (REEP-ID) advocates exploiting the abstract structure of a target domain as the foundation for building cohesive diagrammatic representations. Previous research explored the application of this approach to the display and optimisation of solutions to complex, data rich, real world problems with promising results. This paper demonstrates the application of these principles to generate interactive visualisations for solving complex combinatorial optimisation problems, in this case the University Exam Timetabling Problem (ETP). Using the ETP as an example the principles of REEP-ID are applied, illustrating the design process and advantages of this methodology. This led to the implementation of the VAST (Visual Analysis and Scheduling for Timetables) application, enabling individuals to solve complete instances of the ETP using interactive visualisations. Rather than using automated heuristics or algorithms, VAST relies entirely on the user's problem solving abilities, applying their knowledge and perceptiveness to the interactive visualisations maintained by the computer. Results from an evaluation of VAST support the use of the REEP-ID methodology and the case for further research. In the closing discussion these findings are summarised together with implications for future designers.