Project Title: A Lightweight Simulated Annealing-Based Scheduling System for 
Amateur Sports Competitions with Mobile Deployment and Dynamic Adjustment   
Student: Yuqing Liu
Course: MS in Computer Science (Artificial Intelligence)  

Abstract: 
This thesis investigates intelligent optimisation techniques for amateur sporting event
scheduling with a emphasis on responsiveness to real-world disruptions and deployability
on mobile devices. Amateur events are prone to sparse venue scheduling, variable time
slots, and higher sensitivity to changes at the eleventh hour, making manual scheduling
tedious and inefficient. For addressing these challenges, the project devises a schedule 
system that combines a lightweight mobile application prototype and metaheuristic optimisation. 
The project starts with a greedy strategy and structural patterns for generating plausible 
initial schedules that are eventually optimised by Simulated Annealing (SA). The project also 
devises a lightweight variant of SA by reducing iterations and invoking a faster cooling 
schedule in an effort to achieve practical efficiency under the resource constraint of 
smartphones. The system also includes dynamic adjustment mechanisms for handling routine 
disruption scenarios such as unavailable venue, conflicting timeslot, or team withdrawal. 
The algorithm in such cases performs partial rescheduling or whole re-optimization according 
to the level of intensity of the disruption. The mobile prototype has been developed using
Python and Kivy, and a light interface has also been provided for users for competition
parameter input and viewing schedules being produced. Evaluation covers experiments on 
artificial data and test instances from the International Timetabling Competition on Sports 
Timetabling (ITC2021). Performance indicates that the lightweight SA trades some optimisation 
depth for a significant decrease in runtime and a decrease in memory consumption, to make it 
executable on a mobile. Dynamic adjustment experiments also illustrate that partial repair 
strategies are able to realise a lower rate of disruption than that for full rescheduling, 
yet with competitive quality for the solutions. Overall, we demonstrate the practicability 
for deploying intelligent optimisation techniques for amateur sports scheduling on handsets 
and display a tradeoff between algorithm quality and operational efficiency. The findings 
reveal insights on lightweight optimisation design, disruption handling, and scheduling engine 
integration within mobile applications.