Data and code for "Weather year outweighs algorithm choice in hybrid renewable system sizing"

This deposit contains the input data, generated results and analysis code for the study "Weather year outweighs algorithm choice in hybrid renewable system sizing". Contents:raw/ — 228 cached hourly PVGIS-ERA5 site-year series covering 12 Indian sites over 2005 to 2023, downloaded from PVGIS v5.3 at slope 0 degrees, azimuth 0 degrees and 14 per cent system loss.results/ — 748 NumPy archives holding the 684 main benchmark blocks, 12 convergence blocks, 12 evaluation-budget blocks, 28 five-variable blocks and 12 multi-year reliability matrices.tables/ — the derived CSV tables behind every table and figure in the paper.figures/ — the published figures.Python source for the system model, benchmark, analysis, reference repair and verification, together with manifest.json recording every parameter and library version, the run logs, and REPRODUCE.md. The study benchmarks seven metaheuristics and a random-search control over 684 site-year blocks and 179,520 optimisation runs against reference optima located by dense enumeration, paired with a multi-year robust sizing procedure. The three demand profiles are synthetic. Meteorological inputs are reanalysis rather than site measurement. PVGIS places no restrictions on the use of its data.