Investigation Report: Truck–Drone Coordinated Emergency Distribution Path Optimization Under Spatiotemporal Uncertainty of Post-Disaster Road Conditions

Academic fraud report · Geng Detector

Summary

Verdict: Highly suspicious, recommended for in-depth investigation. This Chinese review identifies two principal data-integrity concerns in a routing-optimization paper published in Chinese Journal of Management Science (DOI: 10.16381/j.cnki.issn1003-207x.2025.0834). First, reported variances across multiple experimental groups (Tables 7–10) appear to be integer multiples of a single base value V₀ = 1.43×10⁻⁴. Examples include ratios of ~9.99×, ~15.01×, and ~19.07×, which is implausible for genuine stochastic optimization runs. Second, ten repeated CPU times for each problem size (Table 6, Ni=50/100/200) cluster within a ~1.2-second band, inconsistent with normal system-level variance. A minor bibliographic issue (misspelled author name in reference [10]: "Liu Q Y" instead of "Y Q") is also flagged. Image-based checks could not be performed because figures were not supplied. Confidence in the variance-ratio anomaly is moderate-to-high; CPU-time anomaly is suggestive but not conclusive. Final determination requires original logs, source code, and raw outputs.

Verdict

🟠 Highly Suspicious — Recommend deep investigation.

Key findings

Multiple reported variances (Tables 7–10) are suspiciously close to integer multiples of a single base value V₀ = 1.43×10⁻⁴.
CPU times for 10 repeated runs at each scale (Table 6) are unnaturally tight, suggesting possible manual fabrication.
Reference [10] contains a Chinese-name romanization error ("Liu Q Y" instead of the expected "Liu Y Q"), indicating careless citation handling.
Image and figure duplication checks were not feasible due to absence of source figures.

Evidence highlights

Variance-ratio anomaly (proposed base V₀ = 1.43×10⁻⁴):
Table 7, truck-only: reported variance 2.161×10⁻³ → ratio ≈ 15.11
Table 8, spatial-only uncertainty: reported variance 2.146×10⁻³ → ratio ≈ 15.01
Table 8, neither dimension: reported variance 2.728×10⁻³ → ratio ≈ 19.07
Table 9, spatial-deterministic/time-uncertain: reported variance 1.41×10⁻³ → ratio ≈ 9.86
Table 10, no variable neighborhood search: reported variance 2.344×10⁻³ → ratio ≈ 16.39
Table 10, Ant Colony: reported variance 1.429×10⁻³ → ratio ≈ 9.993
Table 10, Particle Swarm: reported variance 2.821×10⁻³ → ratio ≈ 19.72
CPU-time anomaly (Table 6, Ni=50): 22.06, 22.39, 21.60, 21.63, 22.08, 21.83, 21.19, 21.76, 21.43, 22.40 s — range = 1.21 s across ten replications.
Citation error: Reference [10] romanizes "刘艳秋" as "Liu Q Y" instead of "Liu Y Q".

Notes

Confidence in the variance-ratio pattern is moderate-to-high; identical integer multipliers across heterogeneous experimental settings are unlikely under genuine stochastic simulation.
The CPU-time clustering is suggestive but not conclusive: tight timings can occasionally arise in deterministic hardware environments, though rarely at the precision shown.
Image-based integrity checks (figure reuse, splicing, blot-style manipulation) could not be performed because the figures were not included in the submitted materials.
Authors should be requested to supply (i) original algorithm log files, (ii) MATLAB/C++ source code used for Tables 6–10, and (iii) raw per-run output arrays so that variance and runtime distributions can be independently reproduced.
All findings remain allegations until verified by the journal or institutional investigation; innocent explanations (e.g., shared pseudo-random seed combined with analytic scaling) must be formally ruled out.