back to main page

Results of the GECCO'2017 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 17 participants in the field. The best function value per problem and participant (1000 times 17 double precision numbers) is listed in this text file.

Participant Ranking

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 nbelkhir AS-AC-CMA-ES Per Instance Algorithm Selection for low dimension and Per Instance Algorithm Configuration of CMA-ES based on problem features extracted a uniform sample of the optimization problem. link 1068.28 5257
2 LB aDTS-CMA-ES Doubly trained surrogate CMA-ES (DTS-CMA-ES) with basic self-adaptation of the number of original-evaluated points per generation. Switch to standard IPOP-CMA-ES after roughly 4 days of computation, or sooner if the run ecountered BBComp server breakdown. CMA-ES tries to restart from unexplored regions. 857.317 5506
3 Simon Wessing Two-stage algorithms link 712.951 5582
4 bujok IDEbdQ
  1. "IDE" - DE with individual dependent parameter settings
  2. "b" - Enhanced mutation with the best base point in the (second) exploitation stage
  3. "d" - diversity control based on the standard deviation of the points in the population (population size is changed inside of Nmin and Nmax boundaries)
  4. "Q" - Nmax boundary from population size update is decreased by half in each quarter of search process
682.493 5800
5 radka 641.706 5940
6 Poly Montreal 562.877 7721
7 Artelys Artelys Knitro Artelys Knitro used in derivative-free mode with multistart link link 521.683 8017
8 anonymous 487.364 7273
9 djagodzi DES - Differential Evolution Strategy 339.278 8083
10 Ralf S. Mix of PSO and GA 279.05 9382
11 Al Jimenez 270.072 9382
12 anonymous953 266.485 9494
13 jarabas CMADE - Covariance Matrix Adaptation Differential Evolution 203.632 9930
14 EMAGIN's Tomcat Emagin's Tomcat (Sparrow mode) - Developed by Mohammadamin Jahanpour 74.9573 11887
15 SebastianGer Initial LHC-based exploitation followed by several short (1+1)-EA runs and pure (1+1)-EA on best result 27.9964 14791
16 Jeremy M Custom algorithm 19.8681 14764
17 bAIz Adaptive Wind Driven Optimization Adaptive Wind Driven Optimization (AWDO) is upgraded version of the Wind Driven Optimization (WDO), where the inherent parameters of the WDO are adaptively selected by CMA-ES method at each iteration. The WDO algorithm is population based nature-inspired algorithm inspired by the atmospheric dynamics equations driving the motion of the wind. The utilization of the actual physical equations makes the algorithm both realistic in nature and very efficient. Check out the homepage at link link 14.1716 14050

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.