 |
High school timetabling is a field of research with only scattered results; most papers in the literature are devoted to a specific school system, or to just one or a few specific schools. For this reason a group of researchers started to work together, first to define a common (XML) format, and secondly to provide datasets from their research in this format, called XHSTT, containing 21 datasets from 8 different countries. The archive XHSTT-2012 with these datasets will be officially released on January 1, 2012. XHSTT-2012 is used for the first round of the competition, while a subset of XHSTT-2012, which will be called XHSTT-ITC2011, will be used for the second round.
The competition consisted of three rounds. In the first round solutions to (separate) instances in XHSTT-2012 could be submitted from January 1, 2012 to May71, 2012. These solutions could be generated by any technique, any resources without any time limits. The best solution received before 7 May 2012 was rewarded with 25 Euros, if it improved the instance contributor's solution. This prize was provided by the research institute CTIT.
The second round of the competition worked with the instances in XHSTT-ITC2011 and hidden instances, that is, instances which were not be publicly released until after submissions close. Most of them were be contributed by the contributors to XHSTT-2012; hence one could expect that the hidden instances would contain "similar" problems to the ones in XHSTT-ITC2011. The constraint types ("AvoidSplitAssignmentsConstraints" and "LimitWorkloadConstraints") were excluded from this round of the competition. A separate XML (input)file was provided for each instance and the participant's program should generate an XML (output)file containing the instance as it was, with a solution to this instance. The input filename and the output filename, the calculation time (in seconds) along with a randseed are parameters to the command line.
Participants had to benchmark their programs and run the program on the instances in XHSTT-ITC2011. The results were submitted at May 7, 2012. Based on the submitted results, the 4 finalists were be selected. These finalists were requested to send their executables (expecting the parameters as described above) to the organizers. The organizers ran the participants' programs 10 times (with different randseeds), and for each instance the solutions were ranked, leading to a rank score from 1 (highest) to 5 (lowest). The finalist with the highest average rank score on the hidden instances wins the second round of the competition. This winner received 500 British Pounds as well as a refund of the conference fee at PATAT 2012. This prize was provided by PATAT.
At June 11, 2012 the hidden instances were archived and released to XHSTT-ITC2011-hidden. For the third round participants could submit a solution to the full archive XHSTT-ITC2011-hidden till July 11, 2012. Here again there were no limits on technology used. The ordering is in the same way as in round 2, by rank score. This winner received 500 Euros. This prize was provided by EventMAP.
For general information about the competition and instances, please contact Gerhard Post (g.f.post at utwente.nl). For more specific question, please have a look at organisers.