We ask authors to submit papers describing their systems in PDF format before February 11th 2011. Papers should be formatted in accordance with the AAAI style template and may be at most 8 pages long. For formatting instructions visit this url. Please note that the learning track of IPC-2011 will be held in parallel to the ICAPS 2011 Workshop on planning and learning. The workshop organizers strongly encourage participants of the learning track to submit papers describing the ideas behind their systems.
The submission deadline is January 20th 2011. We follow the ICAPS rules regarding time zones: Your submission is in time as long as there is some time zone in which it is still January 10th. In other words, the deadline is 11:59 PM UTC-10 (Honolulu time). The deadline is strict. No extensions will be granted.
Systems are submitted by email to Sergio Jiménez <sergio.jimenez.celorrio AT gmail DOT com>:
The learner and planner programs must be submitted as a compressed archive (zip, tar.gz or tar.bz2).
- The compressed archive must include the full source code of the learner and planner programs to be published on this website after the competition. It must not contain any executables.
The compressed archive should not contain any unnecessary files (editor backup files, .CVS or .svn directories, .DS_Store files, object files, bytecode, ...)
- README files that may help with trouble-shooting the planner are appreciated.
We require that everyone uses the following command-line format:
The learning stage
learner -o <domain file> -t <training folder> -k <DCK folder>
The three options are standard. First option specifies the path to the domain definition file. Second option specifies the path to the folder with the set of training problems. Third option indicates the output folder for the learned Domain-specific Control Knowledge (DCK)
The testing stage
planner -o <domain file> -f <problem file> -p <plan file> -k <DCK folder>
The first three options are standard. The <DCK folder> option is optional and will only be specified when evaluating the effectiveness of the learned knowledge. When this option is not specified, the planner should run in knowledge-free mode. The <plan file> option indicates the filename in which the result plan should be stored. If your planner produces multiple plans, please don't reuse the same filename for the generated plans, because this will lead to problems if your planner times out in the middle of writing a new plan. Instead, append .1 to the plan filename for the first plan that is generated, .2 for the second, and so on. We will evaluate the last complete plan that was generated, so please only output plans in increasing order of quality. (Of course, there is no point in producing lower-quality plans than ones that were previously generated anyway.)
The solution should be written to the result file in a format understood by the VAL validator Example of a solution plan for a Blocksworld problem
Compilation and Execution
Each system submission should contain a shell script, named build (note the letter case), which completely builds your learner and planner programs. Be sure that your build script is executable. You may assume that build is run from the directory in which it resides. In the common case that you want to use the make tool to build your learner and planner programs, your build script should look like this:
#! /bin/bash make
- Your learner and planner programs may write whatever they want to the stdout and stderr streams. Diagnostic output to these streams will be logged, and we encourage you to produce any output that may help in troubleshooting the planner.
- Your learner and planner programs will be run with limited user rights, but still please make sure that they doesn't contain any operations that can wreak havoc on the computer. In particular, they must not write to any directories outside the directory it is run in (creating and using subdirectories is fine), and it must not use the network.
If your programs generate any temporary files, we will automatically clean these up after each planner run, restoring the planner directory to its previous state. Don't create temporary files with names ending with .pddl, .soln or .log, as we will use such names for inputs, outputs, and redirected stdout/stderr streams, respectively.
- If your learner or planner programs use randomized algorithms, please initialize the random seed to a fixed constant such as 2011. If there are any reasons to expect that your planner won't generate reproducible results, please tell us clearly in the submission email.
The testing stage will be run in a cluster of 11 nodes. Each node is an INTEL XEON 2.93 GHZ QUAD CORE processor (64 bits) using Linux. Up to 4 GB of RAM memory and 750 GB of hard disk will be available for each planner. Each planner will be run in a single node and no planner will be allowed to run in more than one simultaneously. Memory and time will be externally limited.
If you are interested in the learning part of IPC-2011, we strongly encourage you to register for the Google group "ipc2011-learning". The group is intended to be used for communication between competition organizers, potential participants and all other interested parties exclusively for the learning part, on topics including but not limited to:
- Announcements from the competition organizers
- Technical questions surrounding the competition (e.g. required level of PDDL support)
- Competition rules (requirement to publish sources, scoring system, etc.)
In order to participate in the learning part of the competition, each team should send an email to Sergio Jiménez Celorrio with the registration form completed. As a default rule, we do not allow a team to enter more than two planners, or versions of the same planner. However, depending on the overall number of the registered teams and planners, if a convincing argument can be made that there are significant differences between the planners, more than two planners can be accepted.