Renewable energy forecasting using optimized quantum temporal model based on Ninja optimization algorithm

Abstract

Abstract Artificial intelligence allows improvements in renewable energy systems by increasing efficiency while enhancing reliability and reducing costs. Renewable energy forecasting receives substantial improvement by applying deep learning methods as one of its promising approaches. The research utilizes QTM with NiOA optimization for achieving maximum forecasting performance. NiOA functions through critical optimization processes when enhancing deep learning models with high accuracy for large complex datasets by selecting the most appropriate features. Fundamental data preparation steps, including normalization scaling, and gap handling, play a vital role before using input data for reliable renewable energy forecasting operations. Using the Ninja binary optimization engine produces superior results than all tested binary algorithms, including SBO, bSCA, bFA, bGA, bFEP, bGSA, bDE, bTSH and bBA, resulting in enhanced classification accuracy. The superior capability of bNinja to choose optimal features establishes its usefulness for renewable energy forecasting applications. Experimental implementation revealed that incorporating the Ninja Optimization Algorithm with the QTM model delivered the best R 2 performance at 95.15% with an exceptional RMSE value of 0.00003, thus establishing its ability to optimize renewable energy forecasting accuracy.