Overall optimization of distribution networks

Network reconfiguration and capacitor switching are important measures to reduce power loss and im prove security and economy in automation of distribution. A new method based on parallel genetic algorithm is proposed to search the whole problem space for better solution. Multiple populations evolve independently and communicate periodically, which simulates parallel computing process to save computing time. The results show that the method is robust and has better benefit than the alterafive iteration method. In addition, the effect of overall optimization is better than optimization alone. Power loss can be reduced and the level of voltage can begreatly improved.