Abstract: Research on distributed generation in power systems is of great interest due to its potential benefits in reducing environmental impact and improving overall system reliability and efficiency. The switch allocation problem (SAP) encompasses a series of decision-making faced by power distribution utilities concerning: (i) the number; (ii) type (manual or remote-controlled); (iii) capacity; and (iv) location of switches in a network to minimize operational costs while maintaining acceptable levels of reliability. This work proposes an efficient solution methodology for SAP that can solve large networks with distributed generation. The methodology is based on memetic algorithms embedded with a hierarchical population and local search procedures. Statistical analysis is conducted to determine the optimal values of hyperparameters. A case study on a real-life network (5523 nodes) from the state of São Paulo (Brazil) shows the methodology providing substantial cost reductions compared to the real-world solution implemented by the utility, and the potential benefits resulting from the presence of distributed generation. Also, the proposed approach was evaluated with six well-known instances present in the literature. The proposed methodology solved most of these instances in under minutes and achieved a substantial reduction of the reliability index metric employed, namely, SAIDI Ver menos