Multi-Timescale Edge–Cloud Control under Correlated Batch Demand and Endogenous Degradation: Structural Analysis and Approximate Optimization

This dataset provides the reproducibility data, manuscript source, publication figures, and verification code supporting the manuscript “Multi-Timescale Edge–Cloud Control under Correlated Batch Demand and Endogenous Degradation: Structural Analysis and Approximate Optimization.” The archive contains the V1.19 LaTeX source, bibliography, reference PDF, 12 frozen CSV ledgers, 11 vector figure files, editable TikZ source for the state-feedback system diagram, Python source code, pinned environment information, SHA-256 checksums, and automated verification scripts.

The computational materials cover the equal-mean Poisson, MMPP, and BMAP comparisons; the structural-regime map; the verified batch-size-triggered edge-admission instance; the correlation-induced policy-persistence campaign; the approximation-architecture and algorithmic ablations; the hysteresis analysis; and the state-space and formulation-size audit. The core reproduction script independently regenerates the equal-mean, structural-regime, and N = 20, 40, and 80 scaling campaigns in a separate output directory without overwriting the archived results.