Enhancing generalization in genetic programming hyper-heuristics through mini-batch sampling strategies for dynamic workflow scheduling

Yifan Yang, Aaron Chen, Hui Ma, Sven Hartmann, Mengjie Zhang

Abstract: Highlights•GPHH algorithm with instance sampling strategy fills critical research gap.•Mini-batch strategies outperform rotation, enhancing generalization.•Hybrid strategies further boost generalization ability of GPHH.•GPHH with hybrid strategies excels at generalizing to large-scale problem instances.•Population diversity analysis enriches GPHH understanding.