Multi-fidelity Constrained Optimization for Stochastic Black Box Simulators

TL;DR

This paper presents Scout-Nd, a novel algorithm for efficient constrained optimization of stochastic black-box simulators using gradient estimation, noise reduction, and multi-fidelity schemes to improve performance and reduce computational costs.

Contribution

The paper introduces Scout-Nd, a new method combining gradient estimation, noise reduction, and multi-fidelity approaches for high-dimensional, stochastic, black-box constrained optimization.

Findings

Scout-Nd outperforms existing methods on benchmark problems.

The approach reduces computational effort significantly.

Effective gradient estimation improves optimization accuracy.

Abstract

Constrained optimization of the parameters of a simulator plays a crucial role in a design process. These problems become challenging when the simulator is stochastic, computationally expensive, and the parameter space is high-dimensional. One can efficiently perform optimization only by utilizing the gradient with respect to the parameters, but these gradients are unavailable in many legacy, black-box codes. We introduce the algorithm Scout-Nd (Stochastic Constrained Optimization for N dimensions) to tackle the issues mentioned earlier by efficiently estimating the gradient, reducing the noise of the gradient estimator, and applying multi-fidelity schemes to further reduce computational effort. We validate our approach on standard benchmarks, demonstrating its effectiveness in optimizing parameters highlighting better performance compared to existing methods.