Project Title: Automated Scheduling of Sport Competitions on Multiple Venues
Student: Jonathan Marston
Course: BSc Hons Computer Science	

Abstract:
The aim of this project was to generate a heuristic scheduling method 
that could be used to solve the problem of scheduling sports competitions 
on multiple venues described by Urban & Russell (2003) and a similar 
problem occurring in the University of Nottingham (Section 1.2). I 
developed an initial feasible solution algorithm that generates complete 
feasible schedules where each one team plays only once on each day and 
each team plays at all times. This algorithm developed here has also 
been shown to work for up to 50 competitors whereas the method by 
Urban & Russell was applied for up to 16 competitors. Following this 
initial feasible solution algorithm, heuristic methods were designed 
to improve the quality of the solution while satisfying a number of 
constraints. For problem instances 4 - 50 competitors the heuristics 
obtained solutions with very good quality comparable to those reported 
in the literature and the final schedules produced were very useable. 
To make sure that the initial feasible schedule was not restricting 
the solutions cost function improvement, I generated a infeasible 
starting solution and ran the same heuristic methods which generated 
a much worse infeasible solution. This showed that starting from an 
initial feasible schedule ensures the schedules are feasible but also 
of good quality after heuristics have been run on them.