PGA-Bilevel: A Penalty Genetic Algorithm for KKT-Reformulated Bilevel Programming Problems

Résumé fourni par la source

Python implementation, benchmark suite, experiment runner, and statistical-analysis pipeline accompanying the manuscript 'A Penalty-Based Genetic Algorithm for KKT-Reformulated Bilevel Programming Problems'. Includes the M1-M10 method ablation family, the M11 closed-loop feedback penalty controller, the M12 multiplier-block cooperative-coevolutionary decomposition, the Examples 1-9 / SMD1-12 / SMB benchmark problems, and the raw per-run results (CSV/JSON) behind every table and figure in the paper.