In heteroepitaxial growth experiments, the choice of the substrate is a key step for the realization of high-quality overlayers. This strategy imposes constraints in terms of interface optimization. This optimization can be achieved (i) by selecting a substrate which matches the overlayer properties and (ii) when this matching can not be realized at a high level, by using an optimizing growth process (e.g., buffer layer BL-assisted heteroepitaxy) In this work, we discuss these problems within the framework of a theoretical approach [I] based on the elasticity theory.