soogo.utils module

Utility functions for Soogo.

soogo.utils.find_pareto_front(fx, iStart: int = 0) → list

Find the Pareto front given a set of points in the target space.

Parameters:

• fx – List with n points in the m-dimensional target space.

• iStart (int) – Points from 0 to iStart - 1 are already known to be in the Pareto front. (default: 0)

Return type:

list

Returns:

Indices of the points in the Pareto front.

soogo.utils.gp_expected_improvement(delta, sigma)

Expected Improvement function for a distribution from [1].

Parameters:

• delta – Difference \(f^*_n - \mu_n(x)\), where \(f^*_n\) is the current best function value and \(\mu_n(x)\) is the expected value for \(f(x)\).

• sigma – The standard deviation \(\sigma_n(x)\).

References

[1]

Donald R. Jones, Matthias Schonlau, and William J. Welch. Efficient global optimization of expensive black-box functions. Journal of Global Optimization, 13(4):455–492, 1998.