This paper presents a methodology which determines the optimal allocation of power quality monitors, in order to monitor the occurrence of voltage sags and swells in distribution networks. Initially, the methodology characterizes the system under analysis regarding the occurrence of voltage sag and swells. This characterization is performed through the simulation of several short-circuits at different points of the system being studied and taking in to consideration several conditions (fault impedance, fault type, etc.). For this purpose, a new method that defines the most relevant short-circuit conditions is proposed.After the system's characterization, the methodology makes use of Genetic Algorithms (GA) to define the minimum number of monitors required to monitor the whole system, and also the places where these monitors should be installed in the power network.