News Details | HEAL

Combinatorial Black Box Optimization Competition (GECCO'16)

Jul 24, 2016

We are proud to announce that HEAL member Andreas Beham scored 2nd place with his algorithm titled MemPR submitted to the 2nd Combinatorial Black Box Optimization Competition (CBBOC) at the GECCO 2016 conference in Denver, USA. MemPR competed in the track where no parameter tuning was allowed against 5 other algorithms. Only three algorithms could manage to take a lead in a different subset of the problem instances. MemPR was especially successful when applied on Ising Spin Glasses problem instances.

The top three algorithms in the category without parameter tuning were reported to be

  1. CMA-VNS2
  2. MemPR
  3. P3

The ranks were computed using the Schulze method to find the Condorcet winner.

We would like to thank the organizers of doing a great job in organizing the competition, creating the competition framework for a multitude of programming languages and carrying out all the computations. We would like to encourage everyone to consider the competition at next year's GECCO 2017.