EPSRC Funded Research
Much of the following work is either directly or indirectly supported by existing or recent EPSRC awards.
Temporal Planning
A major area of research interest is in temporal planning. Our group has worked on this for some time, with successful advances being shown in LPGP (a Graphplan-based temporal planner) and various versions of CRIKEY.
Planning with Metric Resources
One of the most successful general-purpose heuristics for planning in metric problems is Hoffmann's extension of relaxed planning to metric values: Metric-FF. Unfortunately, this is not effective in a wide range of interesting metric problems where metric quantities are intertwined in some way (the ability to exchange one commodity for another, for example). This happens in even very simple domains, where metric quantities are used to measure carrying capacity and load is exchanged for space during loading or unloading. We have developed a LP-RPG hybrid that is able to solve interesting and complex metric problems that currently cannot be solved by any other planners.
Planning with Continuous Processes
A longstanding interest in the problem of managing processes as part of planning. Processes are continuous causes of change that are initiated by actions of the agent, but then continue independently. They can be stopped by the agent, under certain circumstances, but the key is to interact with the changes they cause, exploiting opportunities and avoiding difficulties.
Planning for Space Applications
I have a well-established research collaboration with SciSys Ltd in Bristol. We are working together to establish the use of planning in space applications for the European Space Agency. Our current project is examining the role of intelligent replanning in the ExoMars mission.
Plan Validation
Plan validation underpins a great deal of the work we are involved in and we have developed a sophisticated system, VAL, that is capable of validating complex plans under a wide range of assumptions.
Symmetry and Near-Symmetry in Planning and Beyond
Symmetry is an area we have explored for some time, but more recently, we have been intrigued by near-symmetry. We have been pursuing some interesting investigations into the role of near-symmetry in SAT solving.