The Research And Application Of The Hybrid Algorithm Based On Machine Learning And Multi-objective Optimized Load Forecasting

Load forecasting plays a significant role in economic and social development.As a challenging but indispensable task,load forecasting has become one of the hot topics in the energy field.The shortcomings of the existing forecasting methods are that they ignore the close relationship between the input and output of the prediction model,so it is difficult to capture input features that are highly related to the output.In addition,the research focuses on how to improve the forecasting accuracy,ignoring the stability of prediction results.Therefore,in this paper,a novel,robust hybrid forecasting system was developed,composed of,four modules: 1)data preprocessing,2)forecasting,3)optimization and 4)evaluation.In the data preprocessing module,an effective data preprocessing scheme based on singular spectrum analysis and grey correlation analysis was used to produce a smoother time series and to mine the best input and output structure for the model.Then,an extreme learning machine optimized using a multi-objective genetic algorithm that considers both forecasting accuracy and stability was employed to realize load forecasting.Furthermore,a generalized regression neural network was trained using the same data set as training the extreme learning machine model to perform forecasting.Finally,to overcome the drawbacks of using single models,a simulated annealing was utilized to optimize the combined parameters.The half-hourly load data from New South Wales and Tasmania were collected as training and test data to complete simulation experiments to verify the proposed model.Experimental results show that the proposed hybrid system obtains more accurate and stable results than traditional prediction models.