An Improved PSO-Optimized RBF Forecasting Algorithm and Its Application

In order to enhance the forecasting capability of the RBF neural network for time series data, a Particle Swarm Optimization (PSO) algorithm is introduced to optimize the neural network parameters. Due to the tendency of the PSO algorithm to fall into local optima when dealing with complex high-dimensional nonlinear problems, this paper proposes an improved adaptive adjustment of the inertia weight IP-R algorithm and applies this algorithm to trajectory prediction in a carbon accumulation environment. Comparative experiments show that the IP-R algorithm reduces the RMSE in trajectory prediction for three types of moving targets by 0.14, 0.15, and 0.16, respectively.