Per instance algorithm configuration of CMA-ES with limited budget

Nacim Belkhir (1, 2) , Johann Dreo (1) , Pierre Savéant (3) , Marc Schoenauer (4, 2)

1 TRT - Thales Research and Technologies [Orsay]
2 TAU - TAckling the Underspecified
3 Thales Research and Technology [Palaiseau]
4 CMAP - Centre de Mathématiques Appliquées - Ecole Polytechnique

Per Instance Algorithm Configuration (PIAC) relies on features that describe problem instances. It builds an Empirical Performance Model (EPM) from a training set made of (instance, parameter configuration) pairs together with the corresponding performance of the algorithm at hand. This paper presents a case study in the continuous black-box optimization domain, using features proposed in the literature. The target algorithm is CMA-ES, and three of its hyper-parameters. Special care is taken to the computational cost of the features. The EPM is learned on the BBOB benchmark, but tested on independent test functions gathered from the optimization literature.The results demonstrate that the proposed approach can outperform the default setting of CMA-ES with as few as 30 or 50 time the problem dimension additional function evaluations for feature computation.

Mots clés: Bio-inspired optimization; • Theory of computation → Continuous optimization; numerical black box optimization; Algorithm Configuration; problem features; CCS CONCEPTS • General and reference → Empirical studies; fitness landscape analysis; empirical study

Domaines: Intelligence artificielle [cs.AI]; Optimisation et contrôle [math.OC]; Machine Learning [stat.ML]

Nacim Belkhir, Johann Dreo, Pierre Savéant, Marc Schoenauer. Per instance algorithm configuration of CMA-ES with limited budget. GECCO 2017 - Proceedings of the Genetic and Evolutionary Computation Conference , Jul 2017, Berlin, Germany. pp. 681-688. ⟨hal-01613753⟩