MachineLearning/bayesian_optimization library

Bayesian Optimization (surrogate-based) - compact

A pragmatic, minimal Bayesian Optimization flow suitable for low-dimensional continuous problems. This module is intentionally focused on readability, testability and engineering hygiene rather than raw performance.

Features:

  • surrogate modeled with a simple Gaussian Process using RBF kernel
  • expected improvement acquisition function
  • sequential sampling with a fixed budget
  • inputs are List<double> vectors (continuous) and objective f(x)->double

Contract:

  • Input: initial sample generator, objective(List
  • Output: best point and value.
  • Errors: throws ArgumentError for dimension mismatch.