Integration of recursive feature elimination in renewable energy sources with evaluation based on relative utility and nonlinear standardization

Abstract

The transition to renewable energy (RE) is essential for sustainable development and energy security, especially in rapidly urbanized areas. This research presents a new decision-making framework designed to identify the most effective renewable energy sources for Shenzhen, China. The framework employs Recursive Feature Elimination (RFE) to condense 14 initial criteria to the 5 most significant factors, thereby improving evaluation efficiency while maintaining analytical accuracy. This method integrates a logarithmic percentage change-driven objective weighting technique (LOPCOW) with evaluation based on relative utility and nonlinear standardization (ERUNS), which is subsequently enhanced through the application of the linear diophantine fuzzy soft-max average (LiDFSMA) operator. This hybrid model addresses uncertainty in multi-criteria decision making (MCDM) by employing advanced weighting and aggregation techniques. Empirical results indicate that solar thermal power is the most effective alternative for Shenzhen, due to its reliable power generation capacity and suitability for areas with direct sunlight. Pseudocode inclusion promotes transparency and facilitates replicability. The proposed method provides a reliable, objective, and flexible instrument for RE planning, significantly influencing the acceleration of sustainable energy adoption in urban environments.