Genetic Programming Dispatching Rules for the Interval Job Shop: Instances, Evolved Rules, Results and Code

Companion data and code for the article "Genetic Programming Hyper-Heuristics for the Job Shop Scheduling Problem with Interval Durations: Robustness-Aware Interpretable Rules that Generalize Across Instance Sizes". The deposit contains the 70 interval Taillard instances and the 12 classical interval instances used as benchmarks; the 220 dispatching rules evolved for the article's experimental arms (main, terminal ablation, robust objective, lambda sweeps and crisp-midpoint control); the primary CSV result files behind every table and figure; and a self-contained Python package (ijsp_gp) implementing the interval arithmetic, the semi-active decoder, the hand- rafted baselines, the GP evolution and the Monte Carlo executional-robustness measure. An equivalence test re-derives the deposited results from the code and data alone.