Improvement of input-output correlations of electric power load forecasting by scatter search

This paper proposes an improving method of input-output correlations of forecasting system for electric power load by scatter search. We have developed electric power load forecasting methods using weather information at Nagoya city. For high accuracy of forecasting, input-output correlations of forecasting system must be improved using weather information of other areas. Weighted average weather information at many cities is utilized for realizing high accuracy of load forecasting. The weight rates are optimized for the highest correlation between weighted average weather information and electric power demands by scatter search. The simulation results reveal the effectiveness of the proposed method