An advanced evolutionary algorithm for load forecasting with the Kalman filter

In this work, the authors design an advanced evolutionary algorithm for optimizing a Kalman filter (KF) load forecasting model. The EA employs parallel architecture and an advanced mutation operator called the "selection follower". Its performance is benchmarked with that of the expectation-maximization (EM) algorithm in minimizing the mean-square-error of the KF prediction. Results show that although the EA requires more function evaluations, it outperforms the EM algorithm consistently.