PhD Opportunity: EPSRC Funded Position
Planning is one of the fundamental problems in Artificial Intelligence. The ability to plan and to reason causally and temporally is one of the key features of intelligent behaviour. Finding optimal solutions to propositional planning problems is PSPACE-hard. However, because of the importance of solution quality in most realistic applications, there is strong research interest in finding novel ways to tackle this complexity. This project addresses the problem of efficient search for optimal plans using the notion of a meta-level constraint satisfaction problem to guide search at the base level. The meta-level representation differs from the standard single-level representation because it captures an understanding of where the hard parts of the base problem lie, supporting intelligent decisions about the selection of variables.
This project concerns the automated identification of meta-variables from the solution structure of search problems. Applicants will have a good mathematical background and strong programming and problem-solving skills. Ideally, they will have familiarity with automated Constraint Satisfaction and Satisfiability methods and AI Planning, but this is not essential as it can be acquired in the first six months of the project. The project is quite theoretical in nature but the results promise to have considerable practical relevance.
The successful applicant will work closely in a small team within one of the largest AI Planning research groups in Europe. The group is internationally well-connected and there are many opportunities for post-doctoral studies around the world. The position is a fully funded post for a home student, including fees and a stipend worth more than 12k tax-free.
Those interested should send a CV and statement of interest to me or Maria Fox. Starting date is intended for summer 2008.