Interested in doing a PhD?


I have a fully funded scholarship available! Please apply!

Who can apply?

Pretty much everyone, but I assume that you are are worth my time, so please don't let me down. That might sound harsh, but I assure you I'm a nice guy. :) Obviously, having attended courses like Artificial Intelligence (search, heuristics, planning etc.) would be invaluable, but also general computer science courses like Complexity Theory / Theory of Computation, or even Graph Theory could suffice. Just convince me that you have what it takes to be successful! I have high expectations from anybody I supervise, but in exchange I offer an for-me-unhealthy amount of supervision and guidance as payback. Also, my field is super-interesting, its community consists of nice people, and there is a lot to explore, so take the opportunity!

Why to apply?

Where to start? :) There are many "soft reasons", like learning how to become a strong independent scientist or just having fun doing cool stuff for a few years :) with me, colleagues from the ANU, my collaborators from various parts of the world, and of course the super-friendly ICAPS community. (ICAPS is the premier conference on planning, the International Conference on Automated Planning and Scheduling, ranked A* by CORE.) ICAPS folk sometimes call themselves ICAPS family, and not without a reason! :)

But you'd probably also like to know what you could do research-wise. I assume that you are already familiar with the most basic concepts of AI Planning. Based on this, on the most abstract level, I'm doing research in two areas: (1) Hierarchical Task Network (HTN) planning and (2) Partial Order Causal Link (POCL) planning. In a very nutshell, the former focuses on a hierarchical approach to planning based on problem decomposition (quite similar to how formal grammars work), while the latter focuses on reasoning about partially ordered sequences of actions. In case you are more interested, you may take a look at this website on hierarchical planning I've put together, and I also recommend my survey paper I published with colleagues at IJCAI 2019. (For POCL planning I don't know of suitable high-level literature that I can recommend.)

In both of these areas, I am interested in:

  • Computational Complexity Investigations
  • Expressivity Investigations
  • Design of Novel Algorithms and Heuristics
  • Verification of Plans and Models
  • Modeling Assistance (creating tools so that "beginners" get guided through the process of creating new models)
  • Planning under Uncertainty (in the context of HTN planning)
  • Planning with time (in the context of HTN planning)
  • Lifted planning, i.e., planning without prior instantiation of all variables to available constants (both in HTN, POCL, and other non-hierarchical planning forms)

Dissertation topics can be selected or created during the process of applying/stating interest (i.e., after a few meetings). That means you will not have to have a topic in mind. We will decide together based on your research interests and capabilities.

How to apply?

First of all, write your email with care, I do not want to see obvious mistakes and carelessness. Don't worry if you are not perfect in writing (or speaking) English, I do not care about this! But I care about the effort you put in. Some people use three or more fonts, which just looks ugly and unprofessional. I will let you decide what to put into your application, but clearly I require a transcript (i.e., mark per course) and a clear statement which knowledge you have about AI planning (including related courses/areas). If you don't do this I know that you did not read this page which proves carelessness -- so I'm not going to accept you. (I am really not rude, in fact I am very altruistic. But if a prospective PhD student doesn't even read my page on how to apply, then clearly that's red flag.) I will also not accept anybody (and most likely not even reply) who lies to me. Sadly, this happens on a regular basis. One applicant after another tells me how inspiring and useful my papers were for his/her research on topic X, where X is a topic that's absolutely obviously completely different from my topic of research (so it was supposed to be flattering but is instead deeply disrespectful as it's a lie in my face). Catching my interest really isn't hard at all: Just be honest, tell me about your background (see above), why you want to work with me, and we are one round further.

I receive applications from all over the world thus with all different kinds of traditions regarding the usage of titles. To create a common ground, let me explain a few rules:

  • If you wish to use a title (such as "Dr." or "Prof. Dr.") then do this in combination with the last (i.e., family) name. In my case that would be "Dear Dr. Bercher". In Australia it is common to use the title "Prof." only after somebody literally acquired this title, i.e., once he or she has the rank "Associate Professor" or "Professor". My rank is currently Senior Lecturer. Although this is equivalent to the Associate Professor in US/Europe, it is still unusual to use that title -- since technically it's not the right one (in Australia). However, "Lecturer" or "Senior Lecturer" are only used as ranks, not as titles, so if you prefer using a title for a Lecturer/Senior Lecturer (like me), using only "Dr." would be most appropriate.
  • If you wish to skip the title(s) but still like to remain formal, then use the last name. In my case "Dear Mr. Bercher".
  • If you wish to communicate on a first-name basis, then only use the first name while skipping all titles. In my case "Dear Pascal". Feel free approaching me on a first-name basis only, I am totally fine with that. Otherwise we can switch to a first name basis after we exchanged a few emails. (Note that scientists are all on a first-name basis among each other. Only with students it's sometimes different because of the age gap. I am however still very happy with a first-name basis; it's all about content, not rank!)

I add all this info since I get so many variants like "Dr. Pascal", which is just wrong. :)

Further information?

There's a bunch of useful webpage/online material:

PhD in planning with another supervisor? (Or maybe just not in Australia?)

If you are a motivated and knowledgeable student, then of course I would love to supervise you! Sadly, there is always limited funding (and time), so I cannot accept everybody. So I list a few options how you could find other positions or supervisors:

  • Here at the ANU I have many colleagues in planning, and many of them might even have scholarships (or positions for post-docs). They might have positions just "for them", others might even have PhD projects where I could join in as a co-supervisor. Either way, here are our announcements!
  • PhD positions in planning from everywhere around the world are regularly posted over the following two mailing lists, so you might consider subscribing them.
    • -- This is a general mailing list for planning.
    • -- This belongs to the ICAPS conference series, the premier conference on Automated Planning and Scheduling.