Project Title: Intelligent Optimisation for a Workforce Scheduler Mobile Application
Student: Abdullah Tukur
Course: BSc Hons Computer Science with Artificial Intelligence	
	
Abstract:
Workforce scheduling is a complex NP-hard optimisation problem with significant practical relevance, particularly for
small to medium-sized businesses that rely on manual methods. This dissertation presents a mobile application for
staff scheduling that integrates reinforcement learning techniques with heuristic optimisation, aiming to produce more
intelligent and adaptable scheduling outcomes. Building upon previous work that implemented Simulated Annealing
in an Android environment, this project introduces a deep Q-network-based framework to dynamically adjust operator
selection, parameter tuning, and worker assignments in real-time.
The system was evaluated across various problem instances, including those in previous dissertations. Results
demonstrate that the learning-based approach slightly improves solution quality and convergence speed. On average,
the learning-enhanced algorithm converged more quickly than the non-learning heuristic baseline, while maintaining
an acceptable runtime on mobile devices. Additionally, user testing revealed app usability and user satisfaction
improvements due to the redesigned interface. These results validate the effectiveness and feasibility of incorporating
learning-driven scheduling onto mobile and other platforms.