Expected Diverse Utility (EDU): Diverse Bayesian Optimization of Expensive Computer Simulators

TL;DR

This paper introduces EDU, a Bayesian optimization method that efficiently finds a diverse set of near-optimal solutions for expensive black-box simulators, enhancing decision-making in complex engineering tasks.

Contribution

The paper proposes a novel EDU acquisition function for Bayesian optimization that explicitly promotes solution diversity within a specified tolerance, with a closed-form expression facilitating efficient computation.

Findings

EDU outperforms existing methods in numerical experiments.

EDU effectively finds diverse near-optimal solutions in practical applications.

The method reveals a new exploration-exploitation-diversity trade-off.

Abstract

The optimization of expensive black-box simulators arises in a myriad of modern scientific and engineering applications. Bayesian optimization provides an appealing solution, by leveraging a fitted surrogate model to guide the selection of subsequent simulator evaluations. In practice, however, the objective is often not to obtain a single good solution, but rather a ``basket'' of good solutions from which users can choose for downstream decision-making. This need arises in our motivating application for real-time control of internal combustion engines for flight propulsion, where a diverse set of control strategies is essential for stable flight control. There has been little work on this front for Bayesian optimization. We thus propose a new Expected Diverse Utility (EDU) method that searches for diverse ``$\epsilon$-optimal'' solutions: locally-optimal solutions within a tolerance level $\epsilon > 0$ from a global optimum. We show that EDU yields a closed-form acquisition function under a Gaussian process surrogate model, which facilitates efficient sequential queries via automatic differentiation. This closed form further reveals a novel exploration-exploitation-diversity trade-off, which incorporates the desired diversity property within the well-known exploration-exploitation trade-off. We demonstrate the improvement of EDU over existing methods in a suite of numerical experiments, then explore the EDU in two applications on rover trajectory optimization and engine control for flight propulsion.