Refined CMA-ES for CEC 2026 Bound-Constrained Single-Objective Optimization
Organizations: Institute of Computer Science, Warsaw University of Technology, Warsaw, Poland
Abstract
We present a simplified and parameter-adjusted variant of CMA-ES designed for the CEC 2026 Bound-Constrained Single-Objective Optimization competition. Unlike typical algorithmic developments that increase complexity by adding mechanisms, Refined CMA-ES systematically removes components that do not improve performance, such as the flat-land escape, while retaining standard features that provide measurable benefits. Adjusting key parameters (population size and initial step-size) yields substantial gains, outperforming the baseline CMA-ES and several state-of-the-art competitors. The paper highlights the practical relevance of individual CMA-ES components under the CEC 2026 ranking, which equally weights solution accuracy and convergence speed, demonstrating that simplification combined with improved parameter settings can lead to superior performance.
Figures & tables
| place | algorithm | points |
|---|---|---|
| 1 | R-CMA-ES (ABC) | |
| 2 | AC | |
| 3 | AB | |
| 4 | CMA-ES |
| Function | R-CMA-ES | CMA-ES | Test |
|---|---|---|---|
| 1 | 0.0e+00 | 1.1e-14 | + |
| 2 | 1.0e+02 | 1.0e+02 | = |
| 3 | 1.6e+02 | 1.4e+02 | – |
| 4 | 1.1e+02 | 7.8e+02 | + |
| 5 | 1.0e+02 | 1.9e+02 | + |
| 6 | 1.4e+02 | 2.8e+03 | + |
| place | algorithm | points |
|---|---|---|
| 1 | R-CMA-ES | |
| 2 | L-SRTDE | |
| 3 | RB-IPOP-CMA-ES | |
| 4 | CMA-ES | |
| 5 | EA4EigSimpTowardsIDE |
| 3.32s | 3.66s | 0.1 |
Covariance Matrix Adaptation Evolution Strategy
Covariance Matrix Adaptation Evolution Strategy (CMA-ES) is a derivative-free optimizer that learns search directions and scales for black-box problems.
See all 14 papers