In general, substations are designed with transformers in parallel, operating at half load, to ensure continuity of power supply even if a transformer in the group fails. However, this practice involves high costs, especially with regard to the arrangement of substations. An alternative solution is the use of shared stocks of reserve transformers and mobile substation units, which must serve a group of substations with the same electrical characteristics. In this context, this thesis proposes the temporary use of mobile substation units, which must act to meet the energy demand while the reserve transformer is installed. This strategy increases system reliability and reduces operational and investment costs for electric utilities. However, the success of this approach is intrinsically linked to the proper sizing and location of stocks. For good results to be achieved, it is necessary to size stocks so that there is no unnecessary investment or an insufficient number of equipment, and consider strategic storage points, so that the time taken to move the spare equipment to the point of failure minimizes the associated costs. Thus, this thesis presents two optimization methodologies based on Monte Carlo simulation and Genetic Algorithm: one to size the number of spare transformers and mobile substation units per year; and another to properly locate such equipment in substations with storage capacity. Both methods were applied to a group of transformers from two real utilities, aiming at proving the method s ability to find a set of feasible solutions from a technical and economic point of view.