Taking account of the mutual impacts of distributed generation, reactive power and network-configuration planning, a new and more comprehensive distributed-generation-planning problem including simultaneous distributed generation, reactive sources and network configuration planning is defined. By solving this problem the amount of distributed-generation resources (DGRs) and reactive-power sources (RPSs) in selected buses of a distribution system, are determined taking account of their outputs at different load levels, tap positions of voltage regulators (VRs) and the status of sectionalising switches, in order to minimise the cost of power, energy losses and the total required reactive power subjected to a given total of distributed generation. Since the defined planning problem is a combinatorial one, a solution algorithm based on tabu search, that is an efficient heuristic method to a combinatorial-optimisation problem, is proposed. Various memory structures including short-, intermediate- and long-term memories have been implemented in the algorithm. The method has been tested on 33-bus and 69-bus radial distribution systems to demonstrate the performance of the algorithm and to investigate impact of some parameters such as maximum limits on size of DGRs (RPSs), constant or controllable outputs of these sources and different control variables sets on results of system planning. In addition, a novel technique for determining the candidate buses to install active (reactive) sources based on clustering system buses in view of assigning DGRs (RPSs) is presented.