Time-of-use Short-term Load Prediction Model Based on Variable Step Size Optimized HBMO-LS-SVM Method in Day-head Market

The short-term load prediction results are important basis for arranging dispatching plans scientifically, decision-making for competitive bidding of power generation enterprises and users. Firstly, this paper constructs a factor index system of time-of-use short-term load prediction of next day in day-head market, and then builds the HBMO-LS-SVM prediction model. In order to prevent falling into local optimum traps, it optimizes the prediction model based on variable step size, which can improve the convergence speed of prediction as well as prediction accuracy in principle.