My research is in the area of Artificial Intelligence Planning. Given some initial state of the world, a plan says what to do, and when to do it, in order to achieve some goals. Research in AI planning explores how the process of finding a plan can be automated, using a planner. As the number of combinations of actions is large, planning systems must be able to make intelligent decisions in order to solve problems: exploration of all possible combinations of actions is computationally infeasible.

Tell me more...

I am particularly interested in planning for problems where time matters and/or where there are costs and rewards on actions. Sometimes, these are inter-related. Anyone who has bought a train ticket to London will know that the time of day matters: it affects the cost of the ticket, and the transport options available. We could find a plan by ignoring time, just making sure the legs of the journey are connected, but it wouldn't be a great plan: we might have to wait a long time for a connection, or pay a high fare. Thus, for automated planning to really be useful here, it has to take time (and costs) into account.

So what is a planner?

A planner is a piece of software. It takes a description of the current state of the world; the actions that could be taken; and the goals that need to be met. Using a combination of techniques, it then searches for a plan, by considering various sequences of actions. Most of my work is on developing anytime planners that try to find some plan quickly; then carry on searching to find a better one.

Research projects

COHERENT (2021–2024)

CHIST-ERA project, funded via EPSRC. Role: principal investigator at King's.

For robots to build trustable interactions with users two aspects will be crucial during the next decade. First, the ability to produce explainable decisions combining reasons from all the levels of the robotic architecture from low to high level; and second, to be able to effectively communicate such decisions and re-plan according to new user inputs in real-time along with the execution.

COHERENT will develop a novel framework to coordinate explanations originated at the different robotic levels and to be able to deliver these explanations during the execution of the task. To have effective interactions, an interface of communication with the user will be developed to both explain and receive inputs in the form of user preferences, requirements or suggestions to execute a task, at different levels of human expertise. Validation will entail a new benchmark to assess acceptance and effectiveness of explanations based on experiments with subjects.

THuMP (2018–2022)

EPSRC project. Role: co-investigator.

Interaction with machines is commonplace in the modern world, for a wide range of everyday tasks like making coffee, copying documents or driving to work. Forty years ago, these machines existed but were not automated or intelligent. Today, they all have computers embedded in them and can be programmed with advanced functionality beyond the mechanical jobs they performed two generations ago. Tomorrow, they will be talking to each other: my calendar will tell my coffee maker when to have my cuppa ready so that I can arrive at work on time for my first meeting; my satnav will tell my calendar how much time my autonomous car needs to make that journey given traffic and weather conditions; and my office copier will have documents ready to distribute at the meeting when I arrive in the office. And they will all be talking to me: I could request the coffee maker to produce herbal tea because I had too much coffee yesterday; and the copier could remind me that our office is (still) trying to go paperless and wouldn't I prefer to email the documents to meeting attendees instead of killing another tree?

This scenario will not be possible without three key features: an automated planner that coordinates between the various activities that need to be performed, determining where there are dependencies between tasks (e.g., don't drive to the office until I get in the car with my hot drink); a high level of trust between me and this intelligent system that helps organise the mundane actions in my life; and the ability for me to converse with the system and make joint decisions about these actions. Advancing the state-of-the-art in trustworthy, intelligent planning and decision support to realise these critical features lies at the centre of the research proposed by this Trust in Human-Machine Partnerships (THuMP) project.

THuMP will move us toward this future by following three avenues of investigation. First, we will introduce innovative techniques to the artificial intelligence (AI) community through a novel, intra-disciplinary strategy that brings computational argumentation and provenance to AI Planning. Second, we will take human-AI collaboration to the next level, through an exciting, inter-disciplinary approach that unites human-agent interaction and information visualisation to AI Planning. Finally, we will progress the relationship between Technology and Law through a bold, multi-disciplinary approach that links legal and ethics research with new and improved AI Planning.

ADE: Autonomous Decision Making in Very Long Traverses (2019-2021)

ERGO: European Robotic Goal-Oriented Autonomous Controller (2016-2019)

Funded under the European Union's H2020 programme. Role: principal investigator at King's.

This pair of successive projects produced a software framework for the development of highly autonomous space robotics missions. In these a robot system, given a high level goal, will (re)plan, schedule and oversee the execution of elementary actions to attain the goal, considering Time/Space/Resource constraints.


I have written several planners over the years. They are all freely available.


OPTIC is a temporal planner for use in problems where plan cost is determined by preferences or time-dependent goal-collection costs. Such problems arise in a range of interesting situations, from scheduling the delivery of perishable goods, to coordinating order-fulfillment activities in warehouses.

Find out more about OPTIC »


The award-winning planner POPF was inspired by the idea of searching forwards for a plan from the initial state, whilst minimising the number of ordering constraints added between the actions in the plan. This leads to plans that have greater scope for performing actions in parallel, and consequently, complete in less time.

Find out more about POPF »


COLIN was the first PDDL planner to support (linear) continuous numeric change. A common use for this capability is modelling numeric resource levels such as battery charged that are continuously changed by an action, and need to be managed carefully in order to produce a plan.

Find out more about COLIN »


See my profile on Google Scholar or the King's PURE system, though the former is more complete.

Professional Activities


Event Organisation