The results of the 3rd MISC-live run

The mancoosi project announces the 3rd MISC-live trial run for the MISC competition of solvers for package/component installation and upgrade problems. The motivation and general principles of the MISC competition are outlined here.

The optimisation criteria

This time we had three tracks. The first two are the same as for the MISC 2010 competition. All three tracks used as optimization criterion a lexicographic combination of some simple integer valued utility functions of a solution. In the first two tracks the criterion was fixed, while for the third track the precise criterion was part of the problem instance. A precise definition can be found here.

  • paranoid: we want to answer the user request, minimizing the number of packages removed in the solution, and also the packages changed by the solution;
  • trendy: we want to answer the user request, minimizing the number of packages removed in the solution, minimizing the number of outdated packages in the solution, minimizing the number of package recommendations that are not satisfied, and finally minimizing the number of extra packages installed.
  • user: we want to answer the user request, and look for an optimal solution according to an optimization criterion provided by the user. The criterion is given by a list of utility functions taken from a fixed list of possible functions, together with a polarity (maximize or minimize) for each of them.

Ranking of Solvers

This has changed since MISC 2010: the former two classes FAIL and ABORT have been merged into one. The updated specification of the ranking algorithm can be found here.

The execution environment

is the same as for MISC 2010. Here is the exact description.

Timeline the 3rd MISC-live run

Submission of SolversNovember 26, 2010, 23h59 UT
Announcement of ResultsDecember 10, 2010

Participants

Detailed Results