In practice, various industrial design problems have continuous, discrete, and categorical variables. Our objective is to manage efficiently these different types of variables within a surrogate-based optimization process. In this work, we propose to redefine the notion of distance between the possible values of a categorical variable (named “attributes”), through the concept of “affinity”. The notion of affinities between attributes can be interpreted as a weighted relationship between attributes. These affinities are usually defined based on a physical intuition of the designer. Indeed, affinities are generally implicitly associated to the behavior of one or several outputs that behave(s) similarly for various attributes. In order to study the impact of the use of affinities, numerical results are presented on specific test problems coming from structural and mechanical design frameworks.