Reinforcement learning for maximum power point tracking of photovoltaic systems: systematic review data

Data and analysis code supporting a systematic review of reinforcement-learning methods for maximum power point tracking in photovoltaic systems, presented in Chapter 2 of the doctoral dissertation of M. Abdelmagid (Khalifa University, 2026) and an associated manuscript submitted to Archives of Computational Methods in Engineering. The deposit contains the eligibility criteria and amendments, available search documentation, bibliographic metadata for 622 database query occurrences and 349 screened records, reviewer-level screening decisions, full-text eligibility outcomes for 95 reports, the extraction dictionary, reconciled data for 77 included reports, source annotations, and scripts for reproducing the descriptive corpus analyses. Two reviewers independently screened the records and extracted study data using common eligibility criteria and an extraction dictionary. Screening disagreements were resolved through discussion, with the first author serving as the final arbiter. Database abstracts, keyword fields, and full-text articles are not included. See README.md for the file inventory, data conventions, and instructions for reproducing the analyses.