Photovoltaic (PV) site suitability analysis has become an essential component of renewable energy planning due to the rapid global expansion of solar energy systems and the increasing complexity of land-use, environmental, economic, and infrastructural constraints. Over the past decade, Geographic Information Systems (GIS) integrated with Multi-Criteria Decision-Making (MCDM) techniques have emerged as the dominant methodological framework for identifying optimal PV deployment locations. However, recent advancements in machine learning (ML), explainable artificial intelligence (XAI), clustering techniques, and large language models (LLMs) are beginning to reshape the field toward more data-driven and intelligent spatial decision-making systems. This review provides a comprehensive analysis of PV site suitability studies published between 2016 and 2026, focusing on methodological evolution, criteria selection patterns, reproducibility challenges, and emerging AI-driven approaches. A systematic literature review was conducted using Scopus, Web of Science, and IEEE Xplore databases, resulting in 72 final studies after multi-stage screening. The analysis reveals that AHP-based GIS-MCDM frameworks remain overwhelmingly dominant, while machine learning and hybrid AI approaches are still limited but rapidly emerging. A total of 63 unique suitability criteria were identified, with climatic, infrastructure, and topographic factors representing the most frequently used categories. The review further highlights substantial challenges related to reproducibility, regional variability, data transparency, and expert-driven subjectivity. Recent studies employing explainable ML, unsupervised clustering, and LLM-assisted weighting frameworks demonstrate significant potential for improving adaptability, interpretability, and automation within renewable energy planning. The review concludes that future PV suitability analysis is likely to evolve toward hybrid GeoAI systems integrating GIS, ML, XAI, clustering, and human-centered AI frameworks to support more robust, scalable, and transparent spatial energy planning.
Ranjgar et al. (Fri,) studied this question.