A Data-Driven Characterization of Multiobjective Optimization Algorithms from the PlatEMO Library

File Description README.md Introduction and overview of a project benchmarking PlatEMO multiobjective optimization algorithms on the COCO/BBOB test suite. algorithms.md List of 137 PlatEMO v4.13 algorithms included in the benchmark, along with excluded algorithms and their reasons for exclusion. prompt.txt System prompt for LLM-based analysis of PlatEMO algorithms, containing detailed rules and a questionnaire for classifying algorithm behavior. questionnaire.txt Structured questionnaire with 134 binary questions for characterizing multiobjective optimization algorithms across categories such as initialization, selection, operators, and archiving. NSGA-II input.txt Source code of the NSGA-II algorithm from PlatEMO, used as a reference input for LLM analysis. Voting_Results.csv CSV file containing voting results, likely classification scores for algorithms produced by the LLM questionnaire pipeline. Supplementary Material.pdf PDF appendix with supplementary materials supporting the algorithm benchmarking research. coco_data.zip ZIP file containing data for 129 ZIP files of used algorithms from the PlatEMO library, each with raw COCO benchmark results across 55 bi-objective test functions.