Automated Chaos Scenario Generation for Civil Aviation Passenger Service Systems
Hao Sun, Ping Chen, Shen Zong, Shujuan Han, Zhi Li, Haiyang Zhang, Haiying Sun · Advances in transdisciplinary engineering · 2025
The increasing complexity of civil aviation passenger service systems has heightened the need for robust fault tolerance strategies capable of withstanding real-world operational disturbances. However, traditional manual testing methods are insufficient to uncover hidden failure modes under dynamic and large-scale deployments. In this study, we propose an automated chaos scenario generation framework tailored for civil aviation service ecosystems, integrating stochastic fault injection with a semantic topology model of system components and interdependencies. The framework leverages graph-based propagation heuristics and reinforcement learning–based scenario synthesis to generate diverse, scalable, and high-impact failure chains. Experimental evaluations conducted on a digital twin of a commercial airline service system reveal that our method achieves a 34.6% higher fault coverage and 41.2% faster anomaly detection compared to static rule-based chaos templates. The proposed approach enables engineers to proactively identify brittle dependencies, evaluate system resilience, and implement recovery logic under real-world-like disruption patterns. This work bridges the gap between chaos engineering theory and practical aviation-grade system validation, laying a foundation for resilient service orchestration in next-generation intelligent air travel infrastructures.