The choice of the drive composition is an obligatory stage in designing a product including electric drives. The large variety of the elements composing the electric drive (electronics, driving and kinematic chain) leads to a quasi-unlimited number of possible alternatives. Models and programs make it possible for the engineer to calculate independently the elements of a given composition, but the choice of the composition often remains subjective. We make the assumption that it is possible to automate the choice of an optimal composition. To check this assumption, the criteria, defining this optimum, are expressed and then are expressed in the form of an equation as a function of variable parameters. Finally it is checked that a set of parameters corresponding to the required optimum can be found in an automatic way. To successfully go through of these steps required the understanding of very diverse subjects, in particular: the use of an electric model, the symbolic system evaluation of the smallest root of a polynomial, the important influence of the trajectory and determination of dynainic rigidity. One of the key element resides in the formulation of the characteristics of al1 the elements in terms of admissible torque.