This project started in around 1990 when Tsang looked at the role of problem formulation (now commonly called medelling) in constraint satisfaction. The usefulness of modelling in problem solving is not new. It is considered more an art than science, which is absolutely true. Tsang argued that modelling in constraint satisfaction is more manageable than modelling in general because a constraint satisfaction problem is defined by nothing but the variables, domains and constraints. A paper was submitted to AI Commummications (European Magazine) in the early 90s, which was rejected with the feedback that "there was not much new".
Borrett started his PhD (1994) in modelling in constraint satisfaction. By setting up a formal framework for defining the problem of model selection and defining criteria for evaluating binary constraint models, Borrett's PhD thesis (1998) made a small but significant start towards constraint modelling.
Selected Publications:
This project is partly funded by EPSRC and University of Essex.
page maintained by: Edward Tsang
Constraint Research Group,
Department of Computer Science,
University of Essex
email: edward@essex.ac.uk
phone: +44 206 872774
fax: +44 206 872788
Updated: 18 December 2001