Due to the need of limiting costs and energy consumption, in real deployments of wireless sensor networks the exact positions of only a restricted number of nodes are generally available. Since a lot of applications require to know where all the nodes have been placed, the issue of estimating the locations of the remaining nodes has attracted a lot of interest in the literature. In this paper, we discuss how the localization problem can be solved by using a two-objective evolutionary algorithm which concurrently aims to maximize the localization accuracy and minimize the number of connectivity constraints non-satisfied by the candidate geometry codified in the chromosomes. The proposed approach has been applied to different network configurations and compared in terms of normalized localization error with a state-of-the-art method based on semi-definite programming. The results show that our approach outperforms the compared method in all the configurations.