Skip to main content

hypervolume_mc

Function hypervolume_mc 

Source
pub fn hypervolume_mc(
    pareto_front: &[Vec<f64>],
    reference_point: &[f64],
    n_samples: usize,
    seed: u64,
) -> f64
Expand description

Monte-Carlo hypervolume approximation for arbitrary number of objectives.

Estimates the hypervolume by sampling random points uniformly in the bounding box [ideal_point, reference_point] and checking whether each sample is dominated by at least one point in pareto_front.

For 2-D problems prefer hypervolume_2d which gives an exact result.

§Arguments

  • pareto_front - Objective vectors (each of equal length).
  • reference_point - Upper bound; must have the same length as each point.
  • n_samples - Number of Monte-Carlo samples (higher → more accurate).
  • seed - RNG seed for reproducibility.