Optimizing Solar Panel Systems using Machine Learning and Ant Colony Optimization

Abstract

In this study, the optimization of solar panel systems is investigated using machine learning algorithms and Ant Colony optimization. The performance of ML models, such as Artificial Neural Network, Support Vector Machine, Decision Tree, and Random Forest, is assessed relying on precision, recall, F1 score, and accuracy. Furthermore, the ANN model is combined with ACO for additional optimization. The results of the experimentation indicate that the ANN model demonstrates the best performance with the following scores: 97.86% precision, 96.50% recall, 97.20% F1 score, and 98.00% accuracy. In addition, when ACO is applied, the ANN model focuses its capabilities to predict the pressure hitting on the panel as follows: 98.50% precision, 97.80% recall, 98.15% F1 score, and 98.80%. The given outcomes also reflect the application of confusion matrices to show the classification of each model, and the results indicate that the most effective way to predict the parameters of the solar panel and optimize the system is presented by the combination of ANN and ACO. In such a way, the importance of using machine learning approaches and optimization techniques to optimize the energy generation efficiency in renewable energy systems is indicated. In the future, extensive research should be performed to assess experiments on large-scale solar panel systems using the proposed approaches and algorithms to scale-up sustainability efforts and minimize the environmental impact.