Research on New Energy Consumption Prediction and Access Planning Based on Particle Swarm Optimization and Mixed-integer Linear Programming

With the rapid development of machine learning, it has been applied to solve various problems in social life because of its unique advantages. To address the current new energy consumption prediction and access planning methods of peak shifting capacity is insufficient. Particle swarm optimisation algorithm and mixed integer linear programming algorithm are proposed and applied to the models of new energy consumption prediction and photovoltaic access planning respectively. The new energy consumption prediction model based on particle swarm optimisation is applied for comparative analysis, and the results show that its average prediction accuracy and precision rate are 94.7% and 95.8%, respectively, which are significantly better than the comparison model. The mixed-integer linear programming-based PV access model is used for comparative analysis, and the results show that it is significantly better than the comparison model in terms of comprehensive cost and PV access volume dimensions, which is practical. The above results show that the two models proposed in the study have good performance and practicality, which are conducive to maximising the consumption of new energy and saving costs.