A Solar Power Forecasting Model Using the Whale Optimization Algorithm

This paper presents a Whale Optimization Algorithm (WOA)-based solar power forecasting framework targeting improved prediction accuracy due to the inherent irregularity and nonlinearity of photovoltaic power generation. A comprehensive literature review highlighted the limitations of conventional statistical methods and the dependency of machine learning and deep learning models on appropriate parameter tuning. To overcome these challenges, WOA was integrated with a forecasting model to ideally tune model parameters and improve the generalization capability. Data collection, preprocessing, model development, WOA-based parameter optimization, and rigorous performance evaluation were some of the steps efficiently used in this method. Methodological clarity and reproducibility were ensured through mathematical modeling of WOA, algorithmic implementation, and a detailed flowchart provided. Results demonstrated that the proposed WOA-based model consistently outperformed the baseline models and other optimization techniques, such as Genetic Algorithm (GA) and Particle Swarm Optimization (PSO), in terms of performance efficiency. Performance evaluation using standard metrics, including Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Squared Error (MSE), demonstrated a significant reduction in forecasting errors with the proposed approach. Comparative and convergence analysis showcased that the given method achieves an RMSE of 577.320, an MAE of 459.563, and an MSE of 333298.025, indicating reduced forecasting errors compared to Artificial Neural Networks (ANN), GA-ANN, and PSO-ANN.