In this paper, distributed generation (DG) units are optimally allocated in a network containing unbalanced loads. To do this, parametric and cost functions have been formulated for this problem. A novel technique is used to solve the unbalanced load flow problem group search optimizer (GSO) algorithm is one of the new swarm intelligence algorithms which is modified in this paper and results are investigated. We simulated a real network of Iranian’s north-west power network and IEEE standard test system. Each objective function is configured for one these networks. For each network two allocating scenarios are proposed: one for the number of DGs and the other for the locations of DGs. In case studies, the results from impoved GSO (iGSO) are compared with simple GSO and particle swarm optimization (PSO) algorithms results
