Reproducibility Data and Verification Code for “Endogenous Stochastic Traffic in BMAP/PH/c Queues with Congestion-Experience Memory: Recurrence, Structural Decomposition, and Computation”
Description
This dataset provides the numerical data, verification code, exact-arithmetic certificates, and independent discrete-event simulation (DES) replication outputs supporting the article “Endogenous Stochastic Traffic in BMAP/PH/c Queues with Congestion-Experience Memory: Recurrence, Structural Decomposition, and Computation.” The study develops a BMAP/PH/c multiserver queue in which completion-triggered Congestion-Experience Memory (CEM) endogenously modifies future correlated batch traffic. The repository supports the exact positive-recurrence analysis based on the endogenous saturated customer rate, the decomposition into CEM-regime occupancy redistribution and within-regime BMAP phase reweighting, the closed-form saturated CEM occupancy and monotonicity results, the exact binary-background resolvent reduction, matrix-analytic stationary computations, customer-Palm waiting-time first-passage calculations, stability-reversal analysis, robustness checks, and independent DES cross-validation. All numerical experiments use analytically specified stochastic models and synthetic parameter settings; no external empirical dataset is used. The repository contains machine-readable numerical results, verification scripts, replication-level DES outputs, phase-reweighting stress-test files, and the exact full-continuum certificate supporting the principal numerical and structural findings reported in the article and Online Resource 1. This version synchronizes the repository metadata with the final MCAP manuscript title and three-author article metadata; the audited numerical results are unchanged from the preceding published version.
Files
Steps to reproduce
1. Download all repository files and place them in one working directory. If downloaded as an archive, extract it while preserving the original filenames. 2. Read README.md for an overview of the package and its correspondence with the main manuscript and Online Resource 1. Use DATA_DICTIONARY.md for definitions of CSV variables and columns. 3. Use Paper1_v5.25_Theorem2_Decomposition.csv to reproduce the statewise and aggregate two-channel decomposition of the saturated customer rate into CEM-regime occupancy and BMAP phase-reweighting components. 4. Use Paper1_v5.25_S5_5_Grid_Verification.csv and Paper1_v5.25_S5_5_AllPair_Dominance.csv to reproduce the ordered-kernel, monotonicity, and finite-grid dominance checks. 5. Run python s5_6_fullA_verification_v5.25.py and compare with Paper1_v5.25_S5_6_ClosedForm_Boundary.csv to verify the full 18-state saturated generator, closed-form CEM occupancy, and stability-boundary calculations. 6. Run python s5_7_exact_binary_resolvent_v5.25.py and compare with Paper1_v5.25_S5_7_Exact_Binary_Resolvent.csv to verify the exact reduced binary-background resolvent against the full saturated-chain calculation. 7. Run python s5_8_phase_reweighting_stress_v5.38.py and compare with Paper1_v5.38_S5_8_Phase_Reweighting_Stress.csv to reproduce the stress test in which phase reweighting contributes 26.6% of the total saturated-rate change. 8. Run python corollary5_continuum_certificate_v5.40.py and compare with Corollary5_Continuum_Certificate_v5.40.csv to reproduce the exact-arithmetic full-continuum certificate for Corollary 5. 9. Run python waiting_completion_count_unit_test_v5.25.py to verify the tagged-customer completion-count construction used in the customer-Palm waiting-time analysis. 10. Use Paper1_v5.25_DES_Replication_Results.csv to audit the DES results against the corresponding CTMC quantities in the manuscript and Online Resource 1. 11. Verify file integrity using SHA256SUMS.txt. The numerical/code files retain their original development filenames because later manuscript revisions changed exposition, journal positioning, authorship metadata, and presentation rather than the audited numerical results. 12. No external empirical dataset is required. All experiments use analytically specified stochastic models and synthetic parameter settings.
Institutions
- Sangji UniversityGangwon-do, Wŏnju