Project Title: Rogo - An optimisation problem's solution returned to a mobile device using cloud-computing and SOA
Student: Mark Hopper
Course: Bsc in Computer Science with Artificial Intelligence

Abstract:
As optimisation problems become more widespread in society (for example work-force scheduling) and for research, 
it is important to investigate modern computing devices, techniques and architectures into their suitability for 
the providing of solutions to optimisation problems. In particular this thesis focused on the notions of cloud-computing, 
service oriented architecture (SOA) and Windows Communication Foundation (WCF). With this in mind, the aim of this 
project was to provide a system on a mobile device, which can be utilised for the attempting in solving of a Rogo 
puzzle but which can also provide solutions via two different methods either natively on the device or a server in 
which is accessible by the internet and utilises technologies such as cloud-computing and SOA as previously mentioned.
Within the software developed for the purposes of this thesis, there are several components which include the Android 
mobile application and the WCF Server accessible by the internet utilising SOA and cloud-computing. Within these 
components there are several main features contained within them, including the ability to generate and view solutions 
for Rogo introductory puzzles 1-7 (utilising LP Solve and associated LP models) and allowing the user to attempt to 
solve either pre-defined or generated Rogo puzzles. Furthermore it allows the user to track their high score information 
regarding how they've been doing in solving a given Rogo puzzle and allows them to continue progress from a previously 
saved attempt that hasn't been completed at solving the puzzle. In addition there are also other less essential features 
included such as the ability to modify settings which affect the behaviour of the application, viewing of a series of 
instructions on the rules and how to solve a Rogo puzzle as well as being able to view and update high score information 
related to all Rogo puzzles.
As part of this thesis, an evaluation was done based on two different measures, how well the software met the requirements 
specification detailed within the design section and also the performance of the solution generations via the two different 
methods. During the evaluation, it was clear that solving a Rogo puzzle is particularly difficult natively on a mobile 
device due to its associated characteristics. Furthermore during the analysis of the usage of a cloud-based WCF server 
over natively generating solutions on the device, it was clear that overall the WCF server is more applicable and suitable 
than on the device, due to it generating solutions around 25-30 times quicker. Finally it was discovered during the 
implementation stage of the models for the generating of solutions, that introductory puzzle 8 was particularly difficult 
to solve due to it being a nine by nine grid and also having no forbidden squares, resulting in a rather large solution 
space and took the solver around 3 and a half hours to find the optimal solution. Additionally with this statistical 
analysis attempts were then made to answer four research questions outlined within the introduction, which were subsequently 
successfully answered and as a result, provided an overview as to the suitability of the technologies utilised within the 
project. In particular how the usage of cloud-computing and SOA amongst other technologies proved more feasible compared to 
generating solutions natively on the device, resulting in around a 25-30 time reduction in solution generation time.