The research reported and discussed in this thesis represents a novel approach to User Interface evaluation and optimisation through cognitive modelling. This is achieved through the development and testing of a toolkit or platform titled Toolkit for Optimisation of Interface System Evolution (TOISE). The research is conducted in two main phases. In phase 1, the Adaptive Control of Thought Rational (ACT-R) cognitive architecture is used to design Simulated Users (SU) models. This allows models of user interaction to be tested on a specific User Interface (UI). In phase 2, an evolutionary algorithm is added and used to evolve and test an optimised solution to User Interface layout based on the original interface design. The thesis presents a technical background, followed by an overview of some applications in their respective fields. The core concepts behind TOISE are introduced through a discussion of the Adaptive Control of Thought “ Rational (ACT-R) architecture with a focus on the ACT-R models that are used to simulate users. The notion of adding a Genetic Algorithm optimiser is introduced and discussed in terms of the feasibility of using simulated users as the basis for automated evaluation to optimise usability. The design and implementation of TOISE is presented and discussed followed by a series of experiments that evaluate the TOISE system. While the research had to address and solve a large number of technical problems the resulting system does demonstrate potential as a platform for automated evaluation and optimisation of user interface layouts. The limitations of the system and the approach are discussed and further work is presented. It is concluded that the research is novel and shows considerable promise in terms of feasibility and potential for optimising layout for enhanced usability.
GOLOVINE, J.C.R.R. 2013. Experimental user interface design toolkit for interaction research (IDTR). Robert Gordon University, PhD thesis.