Previous Research Projects
PhD
Here is a
paper ( (c) Springer-Verlag) that was published at ICCS 2008 on my initial PhD research, and
here is a pre-print version of the journal article in the IEEE Transactions on Services Computing 2009 ( (c) IEEE ).
Integrated Scheduling for Wireless Mesh Networks
The
ISforWMN website details the a project concentrating on scheduling for mesh networks. Two main papers are published from this project
Optimised scheduling for Wireless Mesh Networks using fixed cycle times and
Optimising multi-rate link scheduling for wireless mesh networks you can also access pre-print versions of the text
here ( (c) IEEE ) and
here ( Copyright © 2012 Elsevier B.V )
Naturaly Viewed 3D Display Systems
The NV3D display system is designed as a natural view, autostereo display system which uses up to 36 projectors to piece together different images to different viewpoints located in front of its display screen. The system's screen is an asymmetric diffusion sheet; allowing light to pass through only within a restricted horizontal angle. The NV3D project aims to provide this arrangement as a self calibrating system, allowing additional resources to be added with minimal effort on behalf of the user, and allowing the system to be setup and re-calibrated without extensive user training.
Designing Efficient Routes and Schedule for Public Transport Systems
The project "Designing Efficient Routes and Schedule for Public Transport Systems" focuses on route planning for buses and other public transit systems. Optimizing routes and schedules in public transport systems is a hugely complex problem for which many research studies have recommended more use of automated computer-based tools should be made.
The vast search space and multiple constraints involved with public transport planning make automation exceedingly difficult, as well as computationally expensive, hence the use of HPC resources provides an ideal environment for distributing the computational requirements of a bespoke Evolutionary Algorithm designed to to optimise this problem.
The project concentrates on: the generation of real world problem instances, the development of evolutionary algorithms that will efficiently optimise those problems, and the parallel implementation of those algorithms - enabling the problems of the size required in the real world to be solved. We hope the code developed during the project will eventually be designed towards commercialisation. A
paper on the project was presented at IEEE Congress on Evolutionary Computation, Beijing, China, 6 - 11 July 2014.