The Spatial Load Forecasting Based On Cellular Automata

This paper deeply studies the application of Rough Set and Cellular Automata in spatial load forecasting. It is discovered that using Rough Set in the extraction of Cellular Automata transformation rules, though, can extract the transformation rules from a spatial database which is incomplete, uncertain and within the presence of noise, it is a problem of relatively poor fault tolerance and Anti-jamming capability. So,Neural Network with good Anti-jamming capability and generalization ability is introduced to compensate for the deficiency of Rough Set and improve the convergence speed and accuracy of extracting the transformation rules of Cellular Automata. Based on this, this paper proposes a new spatial load forecasting modal relying on Neural Network as well as Cellular Automata. It also specifically describes its overall design scheme and specific modeling steps. Then, focusing on a study area, we do spatial load forecasting with this new modal and a modal based on Cellular Automata and Rough Set respectively. According to the comparison of two results, it demonstrates this new modal proposed in this paper has a higher accuracy.In the end, based on this new modal and the actual situation and functional requirements of the study area, we classify the modules of spatial load forecast system, analyze and explain specifically each modal using the method of software engineering. Then under the.NET development environment we develop and implement a spatial load forecast system with ArcObjects components and do the spatial load forecast and total load forecast of the study area using this system. Comparing the forecast results with the actual data, it is found that the results are basically identical to the actual situation. It demonstrates this software has a good practical use.