Smart Solar Power Site Selection Through a Hybrid Computational Framework of Neuro‐Fuzzy With Genetic Algorithm Optimization

ABSTRACT This paper presents a new neuro‐fuzzy multi‐criteria decision‐making (MCDM) framework designed to optimize the selection of solar power plant (SPP) sites across India. The framework harnesses the combined strengths of fuzzy logic and artificial neural networks (ANNs), trained using genetic algorithms (GAs). This model evaluates essential factors such as solar irradiance, climate conditions, geographical features, proximity to infrastructure, and air quality. The ANN classifies potential SPP locations based on these criteria, while fuzzy logic handles uncertainties and fine‐tunes the ranking of sites. To enhance interpretability and transparency, SHAP (SHapley Additive exPlanations) is applied to quantify the contribution of each factor, ensuring that the influence of each criterion on the ANN's decisions is clearly understood. The framework's effectiveness is demonstrated through the application of the multi‐layer perceptron (MLP)‐GA model. By aggregating results for each significant criterion using fuzzy computational techniques, the framework generates an overall score for each SPP site. This scoring system ranks the sites according to their suitability, and sensitivity analysis confirms the robustness of these rankings under varying weight distributions. The study focuses on four potential SPP sites in India: Bhadla SPP, Pavagada SPP, West Bengal SPP, and Manipur SPP. The proposed framework effectively identified Bhadla SPP as the most suitable location for solar power development, followed by Pavagada SPP, West Bengal SPP, and Manipur SPP. Comparative analysis with classical MCDM methods, such as AHP, TOPSIS, and Fuzzy VIKOR, showed that the proposed approach maintains consistency while offering stronger discriminative ability and adaptability to complex, uncertain data. This hybrid neuro‐fuzzy‐GA approach integrates both quantitative and qualitative data, providing a robust tool for planners and policymakers involved in the sustainable deployment of renewable energy infrastructure. The model's flexibility allows it to be applied across various regions and energy systems, contributing to the global transition to clean energy. The proposed method is not limited to existing sites in India; it can also be used to identify optimal locations for future SPP developments. Overall, this work demonstrates how hybrid intelligent systems can support transparent, data‐driven, and future‐ready decision‐making in renewable energy planning.