Year 2012

• Self-adaptive Surrogate-Assisted CMA-ES (saACM-ES) usually speeds up the original CMA-ES by a factor between 2 and 4 on uni-modal and moderate multi-modal functions. It preserves several important invariance properties of CMA-ES such as invariance w.r.t. rank-preserving transformations of the objective function and orthogonal transformations of the search space. An important property for practitioners is that only the ranges of hyper-parameters used to build the surrogate should be defined, while the algorithm will find nearly-optimal values of hyper-parameters during the search.
• Download saACM source code in MATLAB (configured for the BBOB framework)
• Download GECCO 2012 paper
• Download bibtex
• If interested, please see the results on noiseless and noisy BBOB'2012 benchmark problems.

A comparison of the proposed saACM algorithms with 40+ classical (BFGS, GLOBAL, NEWOUA,...) and evolutionary optimization algorithms on 20-dimensional problems of the BBOB framework. The figure shows the proportion of functions which can be solved with a given precision (axis y) using a given number of function/model evaluations (axis x).

Year 2013

Comparison of the proposed surrogate-assisted versions of IPOP-CMA-ES algorithms on 20-dimensional Rotated Ellipsoid function. The trajectories show the median of 15 runs.

The proposed algorithms outperform BIPOP-aCMA-ES and demonstrate the best overall performance w.r.t. about 100 optimization algorithms tested during the BBOB-2009, 2010 and 2012. The performance of the proposed HCMA (a hybrid CMA-ES) is comparable to the portfolio 'best2009' algorithm (i.e., an artificial algorithm representing a combination of best results shown by a set of algorithms). See BBOB-2013 workshop paper for details.