Sizing of an electric power generation system requires an analysis of the investment, maintenance, and operation costs. In the case of a generation system that uses renewable sources the sizing it is more complex with regard to a conventional system, due to the randomness of the renewable resources, and to the even high costs of wind generators and photovoltaic modules. This paper presents the optimal sizing of a generation system wind-photovoltaic-fuel cell such that demand of an isolated residential load is met. The function objective is constituted by the costs of the system, and the solution method employed is based on evolutionary computation technique called Particle Swarm Optimizer (PSO). The aim of this work is to minimize the total cost of the system such that demand is met. In order to compare the performance of PSO with other method, the sizing of the renewable generation system is made it also by the heuristic method called Differential Evolution.