Abstract— This paper presents an adaptation of a new meta- heuristic algorithm known as Cuckoo Search (CS) to solve the distribution network reconfiguration problem (DNRC), where the objective is to minimize the real power losses subject to constraints such as voltage levels in nodes, current levels in branches, and maintaining a radial topology in the distribution test system. The performance of the proposed method is compared with a popular meta-heuristic algorithm, Ant Colony Optimization (ACO). The methods have been tested on the IEEE 37 bus radial distribution system. Experiments were carried out to compare the simulation time, which is a critical measure for the real implementation of the method in real power distribution systems. Index Terms— Ant Colony Optimization, Cuckoo Search algorithm, distribution system, distribution network reconfiguration, real power. I. INTRODUCTION istribution networks represent one end of the energy delivery process, in which electric power reaches the final users. They are defined as the part of an electrical system that supplies power from transformation points in the transmission system to the customer. Generally any element from the distribution substation to the customer meter is part of the distribution network. One of the main problems of these networks is the power loss due to Joule effect in distribution feeders. The losses (I ଶ R ) are mostly evidenciable in conductors, and may represent 13% of the total power generation. No wonder reducing active power losses is a priority in power system operation.The benefits of minimizing power losses do not impact only the final users but also energy traders and network operators because active power losses cause increases in marginal energy prices and operating network problems that make them more vulnerable to failure. The latter is important because it can create instability in the system and affect the quality of service in terms of reliability and security. Among the most used techniques for reducing losses in distribution networks, the following can be found: connection/disconnection of capacitor banks [1], [2] , the installation of distributed generation [3], [4], network reconfiguration, and any combination of the previous [5], [6]. The work in this paper will focus in reconfiguration. Due to the variability of the loads in a distribution network, the operation of the system is a complex task. For a fixed configuration of the network, the power losses will not be minimum considering the different kind of loads that vary from time to time. Hence, the reconfiguration of the network becomes a tool for adapting the operation of the system to the varying demand of the users. Reconfiguration is a procedure to handle changes in power system connectivity to redirect the flow of power to the loads across the path that generates less losses to the system. This is accomplished by opening and closing sectionalizers and tie switches. The early work on reconfiguration of distribution systems with the aim of reducing losses was proposed by Merlin and Back [7], using an optimization technique that coupled branches to determine the configuration with minimal losses. In [8], Hong et al use a Multi Agent based Particle Swarm Optimization (PSO) with the objective of minimizing losses in the lines of the system subject to the voltage levels at the nodes and branch current limits. A 33-node distribution network is used as case of study. The same objective function is presented in [9] using Ant Colony Optimization – ACO, tested on a 5- node system. In [10] the meta-heuristics techniques Tabu Search, Evolution Strategies (EN) and Differential Evolution (DE) are used to solve the problem of reconfiguration, adding in the reliability of the system in the optimization formulation. Xin-She Yang & Suash Dev developed in 2009 a nature inspired meta-heuristic algorithm called Cuckoo Search [11]. This algorithm is based in the brood parasitism of some cuckoo species. In addition, this algorithm is enhanced by using Lévy flights rather than by simple isotropic random walks. Studies have shown that CS is potentially more efficient than PSO and genetic algorithms [12]. This paper presents the implementation of a new algorithm for solving the DNRC problem by adapting the Cuckoo Search Efren Herazo, Michell Quintero, John Candelo, Jose Soto and Javier Guerrero. Optimal Power Distribution Network Reconfiguration using Cuckoo Search D