Hybrid RIME Optimization for Robust Feature Selection and Photovoltaic MPPT Under Complex Conditions

Abstract

Photovoltaic (PV) systems rely on maximum power point tracking (MPPT) to ensure optimal performance. The solar system's output characteristic curve has many peaks due to differences in temperature and light intensity, and the conventional MPPT algorithms accomplish unwell in multifaceted and dynamic scenarios. By integrating tent mapping during the initialisation phase, this study augments the algorithm's exploratory competences and presents a hybrid RIME Optimisation method to improve MPPT and feature selection tasks for PV systems operating in partial shading conditions. It also uses piecewise mapping to construct sequences that optimise the algorithm's parameters, so striking a fair balance between local exploitation and global exploration. From the perspective of feature selection, the hybrid approach reduces computational costs while improving classification accuracy by generating optimal subsets using metaheuristics inspired by nature. In the PV MPPT context, the proposed technique accomplishes better in relationship to the tracking speed, stability and accuracy than PSO-BOA, traditional RIME, and IRIME approaches.