STOCHASTIC CONSTRAINT PROGRAMMING FOR MINIMISING RISK
Stochastic Programming is a mature subfield of Operations Research
that adds probabilistic reasoning to Integer Linear Programming in
order to solve multi-stage decision problems. Constraint Programming
is a younger subfield of Artificial Intelligence aimed at
deterministic (non-stochastic) problems, with a rich modelling
language and a large family of powerful solution algorithms.
Stochastic Constraint Programming (SCP) is a recent hybrid of the two
frameworks, designed to compactly model and efficiently solve problems
involving both constraints and uncertainty. However, SCP requires
further development in order to be a useful real-world tool, as most
current solution algorithms do not scale up to large applications.
This project will develop SCP for risk management. It is part of a
project jointly funded by the Irish Research Council for Science and
Engineering Technology (IRCSET) and IBM. Three different strands will
work alongside IBM in carrying out new research to make users better
informed about risk, through a wide variety of mathematical and AI
instruments. Its overarching objective is to bring together
complementary approaches to risk assessment and delivery.
Principal Investigator: Dr S. Prestwich (s [dot] prestwich [at] cs [dot] ucc [dot] ie)
The candidate should have (or be about to obtain) a good honours
degree in Computer Science, Artificial Intelligence, Operations
Research or a related subject.