<![CDATA[A kind of power forecast model combined cellular genetic algorithm with BP neural network was established in this article. Mid-long term power demand in urban areas was done load forecasting and analysis based on material object of the actual power consumption in urban areas of Nanchang. The results show that this method has the characteristic of the minimum training times, the shortest consumption time, the minimum error and the shortest operation time to obtain the best fitting effect. DOI: http://dx.doi.org/10.11591/telkomnika.v11i7.2831]]>
