Short-Term Forecasting of Power Consumption of Power Supply Companies Based on the Integration of Technologies of Analytical, Simulation and Expert Systems

Purpose of research is to improve software support and identify regularities in the processes short-term forecasting power consumption supply companies based on complementary integration data mining models, system dynamics expert systems. Methods. The principles constructing predictive models are given. A analysis has been carried out an ontological model subject area built, taking into account technological market environment. classification methods considered. features information base for forecasting, including actual weather data, have described. requirements making forecasts formulated. block diagram day ahead built modeling software. Results. Scenarios processing Loginom developed using Arimax Neural Network (Regression) processors build meteorological factors. simulation that allows exploring influence factors (temperature, pressure, precipitation) Anylogic. Using Wi!Mi mivar constructor systems, task parametrized; indicators, relationships, rules set; a logical conclusion solution obtained. Conclusion. market's built. It retrospective methods, systems applying Russian Loginom, Anylogic tools.