In this paper, we attempt to improve the use of optimal designs for the technological field by applying Markov Chain Monte Carlo simulations, and by evaluating: i) a hierarchical structure of the observed data; ii) the definition of a specific utility function; iii) model discrimination by the predicting point of view. To this end, the Bayesian T-optimality criterion is used and modified considering the specific aims of the study previously mentioned, and other features, such as the consideration of quantitative as well as categorical variables. The optimization is achieved by exploiting an Inhomogeneous Markov-Chain algorithm and using the R software. The obtained results, considering different simulation scenarios, are satisfactory.