In this thesis, a methodology for the optimal allocation of reclosers in distribution feeders was developed combining the techniques of Chu-Beasley genetic algorithm, for the search of optimal allocation, and Monte Carlo simulation for the reliability assessment. The objective function to be minimized is the non served energy, for what a detailed modeling of the load composition by residential commercial and industrial customers was developed. This methodology seeks the optimal allocation for a number of reclosers given by the analyst because: i. A global solution for the number of reclosers would require huge computational resources, i. e. computer memory capacity and computational time. ii. Due to the limited resources of many utilities, although an optimal number of reclosers is known, there is no the resources neither the capacity to go into loans to buy all them but for some of them. iii. The methodology has to be useful for the case that some reclosers are available to be installed or for checking if the location of existing reclosers is optimal. The analyst has to carefully select the coefficient of variation because, if it is very low the computational time will be very high and if it is very low the optimal solution will be not found.