A Machine Learning Augmented Game Theory-Based Approach to Hybrid Renewable Energy System Optimization

The world’s energy requirements have been on a steady increase; all the while governments have pushed for the shift toward renewable energy (RE). As such, numerous studies have focused on optimizing the design of hybrid renewable energy systems (HRES), integrating passive and controllable sources. However, these studies have been limited in scalability and scope, as the models still focus on specific layouts of HRES, and short-term forecasts for expected energy output. Researchers introduce a novel framework leveraging machine learning to predict HRES yield, addressing scalability and enabling long-term decisions. This novel framework allows decision makers to utilize the concepts of machine learning to process and analyze relevant parameters to predict the expected yield of the HRES components and consider every conventional RE technology in the subsections of the model, namely: Solar, Wind, Hydroelectric, Geothermal, Nuclear, Biomass and Fossil Fuel. After the framework was developed, it was tested on a hypothetical dataset. The framework and resulting model were found to be valid and were further tested through sensitivity analysis.