Probability Distribution of Hypervolume Improvement in Bi-objective Bayesian Optimization

05/11/2022
by   Hao Wang, et al.
11

This work provides the exact expression of the probability distribution of the hypervolume improvement (HVI) for bi-objective generalization of Bayesian optimization. Here, instead of a single-objective improvement, we consider the improvement of the hypervolume indicator concerning the current best approximation of the Pareto front. Gaussian process regression models are trained independently on both objective functions, resulting in a bi-variate separated Gaussian distribution serving as a predictive model for the vector-valued objective function. Some commonly HVI-based acquisition functions (probability of improvement and upper confidence bound) are also leveraged with the help of the exact distribution of HVI. In addition, we show the superior numerical accuracy and efficiency of the exact distribution compared to the commonly used approximation by Monte-Carlo sampling. Finally, we benchmark distribution-leveraged acquisition functions on the widely applied ZDT problem set, demonstrating a significant advantage of using the exact distribution of HVI in multi-objective Bayesian optimization.

READ FULL TEXT
research
02/25/2022

A Robust Multi-Objective Bayesian Optimization Framework Considering Input Uncertainty

Bayesian optimization is a popular tool for data-efficient optimization ...
research
07/08/2018

A Tutorial on Bayesian Optimization

Bayesian optimization is an approach to optimizing objective functions t...
research
11/17/2019

A hierarchical expected improvement method for Bayesian optimization

Expected improvement (EI) is one of the most popular Bayesian optimizati...
research
09/19/2018

A Novel Heat Exchanger Design Method Using a Delayed Rejection Adaptive Metropolis Hasting Algorithm

In this study, a shell-and-tube heat exchanger (STHX) design based on se...
research
01/30/2019

Evaluating Bregman Divergences for Probability Learning from Crowd

The crowdsourcing scenarios are a good example of having a probability d...
research
04/01/2016

A sequential Monte Carlo approach to Thompson sampling for Bayesian optimization

Bayesian optimization through Gaussian process regression is an effectiv...
research
07/13/2017

Distributionally Ambiguous Optimization Techniques in Batch Bayesian Optimization

We propose a novel, theoretically-grounded, acquisition function for bat...

Please sign up or login with your details

Forgot password? Click here to reset