Distribution grids are being transformed from passive to active networks by integrating distributed generation (DG). In this paper, a Strength Pareto Evolutionary Algorithm (SPEA) has been used to estimate the best type of DGs and their optimal capacity and location in a 33-node radial distribution network. In order to improve the voltage stability, expressed by its index, and to reduce the active and reactive power losses, multiple scenarios have been proposed and compared. The backward/forward sweep method of distribution load flow was used to evaluate the three objective functions. The same distribution network has been also used to show the effect of the capacity and location of the DGs on the objective functions.