On sampling methods for costly multi-objective black-box optimization

 

We investigate the impact of different sampling techniques on the performance of multi-objective optimization methods applied to costly black-box optimization problems. Such problems are often solved using an algorithm in which a surrogate model approximates the true objective function and provides predicted objective values at a lower cost. As the surrogate model is based on evaluations of a small number of points, the quality of the initial sample can have a great effect on the overall effectiveness of the optimization. In this study, we demonstrate how various sampling techniques affect the results of applying different optimization algorithms to a set of benchmark problems. Additionally, some recommendations on usage of sampling methods are provided.

Cite

@inbook{sampling-multiobjective-optimization,
  address   = {New York},
  author    = {Ingrida Steponavičė and Mojdeh Shirazi-Manesh and Rob J. Hyndman and Kate Smith-Miles and Laura Villanova},
  booktitle = {Advances in Stochastic and Deterministic Global Optimization},
  doi       = {10.1007/978-3-319-29975-4_15},
  editor    = {P. M. Pardalos and A. Zhigljavsky and J. {\v Z}ilinskas},
  pages     = {273–296},
  publisher = {Springer},
  title     = {On sampling methods for costly multi-objective black-box optimization},
  year      = 2016
}