Short-term Power Load Forecasting System Based On Multi-agent

Load forecasting is the primary gist of production and economic dispatch of powersystem security. Accurate short-term power load forecasting is important for economicarrangement of generating capacity, scheduling of fuel purchases and planning for energytransaction. With the deepened reform of power enterprises, it is necessary to develop aprecise, scientific and intelligent electric load forecasting method.By analyzing the conventional load forecasting process and forecast system, a short-termpower load forecasting method based on multi-agent is proposed. Firstly, the sub-agentdecision-making functions of forecasting system are designed in this method, so as to theintelligent model selection as well as the combination of quantitative and qualitativeforecasting results given by different models are implemented based on the idea of intelligentforecast and combining forecast. Then in the prediction method library, the load forecastingmethods based on IEGHS-LSSVM and ACHS-BP are applied. Due to the special day, themethod of synthesis experts predicting loads is also expanded into the prediction methodlibrary.Finally, in order to verify the proposed method, the relevant historical data is used, andthe simulation results show that the intelligent model provides better accuracy and stabilitycompared with normal models.