1: \begin{abstract}

2: 

3: In the paradigm of computer experiments, the choice of an experimental design is an important issue.

4: When no information is available about the black-box function to be approximated, an exploratory design has to be used.

5: In this context, two dispersion criteria are usually considered: the \textsc{minimax}

6: and the \textsc{maximin} ones. In the case of a hypercube domain, a standard strategy consists

7: of taking the \textsc{maximin} design within the class of Latin hypercube designs.

8: However, in a non hypercube context, it does not make sense to use the Latin hypercube strategy.

9: Moreover, whatever the design is, the black-box function is typically approximated thanks to kernel

10: interpolation. Here, we first provide a theoretical justification to the \textsc{maximin} criterion with respect to kernel interpolations.

11: Then, we propose simulated annealing algorithms to determine \textsc{maximin} designs in any bounded connected domain.

12: We prove the convergence of the different schemes. Finally, the methodology is applied on a challenging real example

13: where the black-blox function describes the behaviour of an aircraft engine.\\

14: 

15: \vspace{0.2cm} \noindent {\bf Keywords:} 

16: computer experiments,

17: kernel interpolation, 

18: Kriging, 

19: \textsc{maximin} designs, 

20: simulated annealing.

21: 

22: \end{abstract}

23: