The $$r$$ r -acyclic chromatic number of planar graphs

Learning to Generate Hard Instances: Towards Robust Solutions for Vehicle Routing Problems

“…Existing deep models can be roughly categorized into two main paradigms [16]: learning-augmented models that adopt neural networks to enhance conventional heuristics, and end-to-end models that learn to construct solutions in sequential or one-shot manner. The former models integrate deep learning with existing domain knowledge of VRPs [17], such as improvement heuristics, large neighborhood search, and specialized solvers [18], [19], [20], [21], [22], [23], [24], [25], [26]. These models usually learn to refine an initial solution iteratively given specific contexts in the heuristics.…”