Load Carbon Rate Forecasting Based on Genetic-Ant Colony Hybrid Algorithm Improved Support Vector Machine

With the implementation of the ‘double carbon’ target, load carbon rate forecasting will become essential primary data for power system operation. However, the sample data of short-term load carbon rate have both volatility and randomness, which challenges load carbon rate forecasting. Population optimization algorithms, especially genetic algorithm and ant colony algorithm are widely used in forecasting. However, conventional genetic and ant colony algorithms can easily lead to local convergence, which leads to a decrease in later search efficiency. This paper uses the genetic-ant colony hybrid algorithm to improve the search efficiency. Firstly, the hybrid algorithm is used to force the current optimal solution to avoid the algorithm falling into local convergence. The improved algorithm optimizes the support vector machine’s key parameters c and g. Finally, the load carbon rate is predicted. The genetic-ant colony hybrid algorithm has a better optimization effect, and the support vector machine improved by the genetic-ant colony hybrid algorithm has a better forecasting effect.