Reliability Assessment of Hybrid Hardware‐Software Fault Systems Based on GSPN Under Mixed Uncertainties

With the increasing complexity and intelligence of system equipment, the likelihood of system failures also rises. Such failures can lead to severe consequences, including property damage and personal injury. Therefore, it is essential to assess system reliability in a reasonable and accurate manner. This paper proposes a reliability analysis method for complex systems that comprehensively considers hybrid failures of software and hardware, mixed uncertainties, and human errors. Specifically, to address the characteristics of hybrid software and hardware failures, Generalized Stochastic Petri Nets (GSPN) models are developed for various failure modes. Markov Chain Monte Carlo (MCMC) sampling and inverse transform sampling are used to simulate cognitive and random uncertainties, respectively. Furthermore, to assess the impact of human errors on system reliability, this paper uses a Proportional Hazards Model (PHM) analysis framework, integrating the Common Performance Condition (CPC) factors set from the Cognitive Reliability and Error Analysis Method technique with the Decision Making Trial and Evaluation Laboratory (DEMATEL)–Analytic Network Process (ANP) method, and constructs a GSPN subnet dedicated to human errors. Additionally, the Monte Carlo method is employed to perform a quantitative analysis of the failure model and determine system reliability. Finally, the accuracy and effectiveness of the proposed method are validated using an equipment module in command‐and‐control systems as a case study.