1: \begin{abstract}

2: The challenge of taking many variables into account in optimization problems may be overcome under the hypothesis of low effective dimensionality.

3: Then, the search of solutions can be reduced to the random embedding of a low dimensional space into the original one, resulting in a more manageable optimization problem.

4: Specifically, in the case of time consuming black-box functions and when the budget of evaluations is severely limited, global optimization with random embeddings appears as a sound alternative to random search. 

5: Yet, in the case of box constraints on the native variables, defining suitable bounds on a low dimensional domain appears to be complex.

6: Indeed, a small search domain does not guarantee to find a solution even under restrictive hypotheses about the function, while a larger one may slow down convergence dramatically.

7: Here we tackle the issue of low-dimensional domain selection based on a detailed study of the properties of the random embedding, giving insight on the aforementioned difficulties. 

8: In particular, we describe a minimal low-dimensional set in correspondence with the embedded search space. 

9: We additionally show that an alternative equivalent embedding procedure yields simultaneously a simpler definition of the low-dimensional minimal set and better properties in practice.        

10: Finally, the performance and robustness gains of the proposed enhancements for Bayesian optimization are illustrated on numerical examples.

11: 

12: \end{abstract}

13: