A Cross-Domain Meta-Analysis of Coherence-Regulated Quantum Algorithms: Unified Evidence from Multi-Benchmark NISQ Studies
Arturo Cerezo, IA Collaborative Panel — Claude (Anthropic) • ChatGPT (OpenAI) • Gemini (Google) • Grok (xAI) · Zenodo (CERN European Organization for Nuclear Research) · 2025
Meta-Analysis — Applied Quantum Computing Track This record presents a unifying meta-analysis of applied benchmarks within the Radial Coherential Dynamics (RCD) research program, synthesizing results across multiple quantum algorithms, system dimensionalities, and problem classes under a single coherence-regulation principle. Building upon the foundational decoherence suppression model (Topological Decoherence Suppression via Coherential Geometry, Zenodo, December 2025, DOI: 10.5281/zenodo.18000641), and its applied benchmarks in variational and optimization algorithms, this work consolidates evidence that a fixed decoherence scaling law T2→T2/αT_2 \rightarrow T_2/\sqrt{\alpha}T2→T2/α (with α≈1.3×10−4\alpha \approx 1.3 \times 10^{-4}α≈1.3×10−4) produces consistent, non-ad-hoc improvements across NISQ-relevant tasks. The meta-analysis integrates four independent benchmarks: (i) VQE for molecular H₄ (0D), (ii) VQE for a graphene-inspired hexagon (2D, PBC), (iii) VQE for a 1D Hubbard chain (OBC), and (iv) QAOA for MaxCut optimization on graphs. All benchmarks employ a controlled A/B methodology in which identical circuits are executed with standard noise (control) and coherence-regulated noise (treatment), with no parameter tuning introduced at the benchmark level. Across these diverse settings, coherence regulation yields stable reductions in energy error (≈ 40–46% for VQE tasks) and optimization gap (≈ 25% for QAOA), with magnitudes consistent with the underlying decoherence scaling and robust to changes in topology, dimensionality, and algorithmic structure. This work does not introduce new physical assumptions or experimental claims. Rather, it provides a consolidated, cross-domain view intended to clarify scope, consistency, and limitations, and to facilitate independent assessment and downstream reuse by the scientific community. All figures, summary data, and benchmark references are provided for transparency and reproducibility.