[alife] PhD Studentship in Social Insect-Inspired Algorithms for Decentralised Control in Distributed Computing (Bristol)

[James Marshall]

PhD Studentship in Social Insect-Inspired Algorithms for Decentralised 
Control in Distributed Computing (Bristol)

Supervising Institutions:
   Department of Computer Science, University of Bristol
   HP Labs, Bristol

Social insect colonies face similar decision problems to those inherent 
in large-scale information processing systems such as the internet, grid 
computing and utility computing centres. These systems are too complex 
for classical control techniques to be applicable, as global information 
on the state of the system is impossible to obtain. However, we can draw 
inspiration from the algorithms employed by social insect colonies. 
These algorithms work robustly using simple rules of thumb, without 
centralised control, and achieve or approximate optimality. Hence these 
algorithms will prove ideal for decentralised optimal control problems 
such as those encountered in distributed computing.

Recently, a new social-insect behaviour has been examined 
experimentally, collective nest-site selection during colony emigration. 
This is a collective decision-making mechanism, that converges on a 
unique solution while exhibiting a tuneable compromise between the speed 
and the accuracy of decision-making. The successful applicant will 
investigate the application of this behaviour to distributed control 
problems in utility computing centres, with input from HP Labs. It is 
expected that the student's research will also lead to developments in 
algorithm design and statistical decision theory.

The successful candidate will have a high quality first degree(s) in a 
numerate discipline such as mathematics, physics or computer science. 
The studentship will be based in the Machine Learning and Biological 
Computation Group, in the Department of Computer Science at the 
University of Bristol. For more information on the Department and the 
Group visit www.cs.bris.ac.uk and 
www.cs.bris.ac.uk/Research/MachineLearning.

The studentship is expected to commence in early October 2008, but an 
immediate start may be possible by negotiation. The studentship will pay 
tuition fees and a stipend of 12,600 GBP per annum for 3 years. UK 
citizens and EU-citizens resident in the UK for at least 3 years 
previously are eligible to apply. Exceptional overseas applicants may 
also be considered.

Apply by email to Dr James Marshall, enclosing PDF copies of your CV and 
a covering letter outlining your interest in the project. It is 
anticipated that interviews with shortlisted candidates will be held 
within 2 weeks of the application deadline.

For informal discussions contact Dr James Marshall 
(www.cs.bris.ac.uk/~marshall)

Closing deadline for applications: 14th March 2008

-- 
James A. R. Marshall
Department of Computer Science
University of Bristol
http://www.cs.bris.ac.uk/~marshall