This research is an exploration of the usefulness of three main optimization algorithms, Genetic Algorithms (GA), Simulated Annealing (SA) and Hill Climbing (HC) in optimizing the layout of the Trail Making Test (TMT)—a popular neuropsychological test that assesses visual attention and task-switching abilities. As cognitive testing quickly shifts to digital, there is a critical need for algorithms that can create digital layouts with precision without affecting its credibility. This research draws a comparison between these three algorithms based on execution time, solution quality and computational efficiency required to identify the most appropriate algorithm for digital TMT layout design.
The experimental analysis was carried out using MATLAB-based implementation tool which evaluated each algorithm’s ability to produce layouts that had minimal edge crossings as well as optimized node distribution. The research also revealed that GA performs significantly better than SA and HC in generating complex valid layouts within short periods of time and with low resource use as stated in the aim of achieving computational efficiency and high-quality layout generation.
Also, this study has led to the development of an open-source API on cloud platform for researchers and practitioners who want to generate optimal TMT layouts thus contributing to cognitive assessment technology at large. It emphasizes the importance associated with proper choice of optimization algorithms for particular applications in cognitive testing while setting a case in point for upcoming studies aimed at integrating advanced algorithmic solutions into psychological test designs. Results suggest further exploration towards hybrid algorithms perhaps combining the speed of SA, HC or even both with quality strong solution provided by GA that are likely to revolutionize digital cognitive health diagnostics.
Keywords: Trail Making Test, Cognitive Assessment, Optimization Algorithms, Simulated Annealing, Hill Climbing, Genetic Algorithms, Test Layout Design.
