Project Title: Developing Android App to Solve Workforce Scheduling Problems
Student: Yusheng Wu
Course: BSc in Computer Science

Abstract: 
As the cost of personnel has become a significant factor in most service-oriented 
companies, the way of arranging staff to finish a group of tasks are crucial in 
today's market. Particularly in a bike sharing company, the manager needs to organise 
the bike technicians to preserve and maintain the bikes in each station every day, 
which is time-consuming if do it manually. Large bikeshare corporations often use 
advanced and expensive software to schedule technicians to reduce maintenance 
consumption and maximise profit. However, there exists a blank area in the market 
for a free and sufficient software that can automatically schedule the technicians 
for small-size bike share companies.
The intention of this project is to develop an android application that would be 
useable for the staffs in a small bike share company, providing various functions 
to managers and technicians to help them improve the work efficiency. In more detail, 
the application focus on solving the technician routing and scheduling problem(TRSP) 
on a mobile platform. The basic idea of TRSP in this project is to assign a set of 
tasks to a group of technicians taken into account skills, locations and time windows. 
A real bike-share company called Bay area is used as the scenario, combining it with 
the features of the TRSP, the system requirements are well informed. Moreover, after 
studying existing technician scheduling algorithm, the Guided Local Search(GLS) is 
chosen to solve this problem. The GLS is a meta-heuristic which sits on top of local 
search algorithms to help them escape the local minima by adding the penalty terms 
in the objective function. After modifying the cost function, whenever the local 
search gets stuck in a local minimum, the penalties are changed and the local search 
will be called again to minimise the augmented cost function. The iterative use of 
local search is an important characteristic of GLS. The main difficulty of the project 
are to quantify the real-world problem into a mathematical model which can be solved
using GLS, and develop an easy-use Android App since achieve a clear and usable design 
is a challenging task. Also, how to adjust the GLS to maximise its ability on the 
mobile platform which has limited resources is another difficulty to be solved.
After widely testing the system and critically evaluate the performance of GLS, the 
built application is proved to be efficient and useful in scheduling the technicians 
for a small-size bike share company. Also, the GLS has a satisfactory performance 
for the technician scheduling in the case of limited resources.