Results of the GECCO'2016 1-OBJ Track
"Raw" Result Data
On each problem participants were judged by the best (lowest) function value achieved within the given budget of function evaluations. There were 14 participants in the field. The best function value per problem and participant (1000 times 14 double precision numbers) is listed in this text file.
Participants were ranked based on aggregated problem-wise ranks (details here and here). The following results table lists participants with overall scores (higher is better) and the sum of ranks over all problems (lower is better) The table can be sorted w.r.t. these criteria.
|rank||participant||method name||method description||software||paper||score||sum of ranks|
|1||Simon Wessing||Two-stage algorithms||
Based on data from a preliminary experiment on test problems, a decision tree was built to decide which two-stage variant to apply on which dimension. The two-stage options were restarted local search and clustering method, the local search algorithms Nelder-Mead and CMA-ES. Before the two-stage algorithm starts, a singular L-BFGS-B run is tried.
|4||Al Jimenez||Curved Trajectories Algorithm (CTA)||link||602.851||6840|
|5||Artelys||Artelys Knitro||Artelys Knitro used in derivative-free mode with multistart||link||link||565.607||6576|
Visualization of Performance Data
The following figure shows an aggregated view on the performance data.
The following figures show the same data, but separately for each problem dimension.