1: \begin{abstract}

2: Many expensive black-box optimisation problems are sensitive to their inputs.

3:   In these problems it makes more sense to locate a region of good designs,

4:   than a single, possibly fragile, optimal design.

5: 

6: Expensive black-box functions can be optimised effectively with Bayesian

7:   optimisation, where a Gaussian process is a popular choice as a prior over

8:   the expensive function. We propose a method for robust optimisation using

9:   Bayesian optimisation to find a region of design space in which the expensive

10:   function's performance is insensitive to the inputs whilst retaining a good

11:   quality. This is achieved by sampling realisations from a Gaussian process

12:   modelling the expensive function, and evaluating the improvement for each

13:   realisation. The expectation of these improvements can be optimised cheaply

14:   with an evolutionary algorithm to determine the next location at which to

15:   evaluate the expensive function. We describe an efficient process to locate

16:   the optimum expected improvement.  We show empirically that evaluating the

17:   expensive function at the location in the candidate sweet spot about which

18:   the model is most uncertain or at random yield best convergence in contrast

19:   to exploitative schemes.

20: 

21: We illustrate our method on six test functions in two, five, and ten

22:   dimensions, and demonstrate that it is able to outperform a state-of-the-art

23:   approach from the literature.

24: \end{abstract}

25: