DRU: Dynamic Modeling of System Safety to Manage Risk and Enable Internal and External Cross-Stakeholder Alignment

传统的因果关系模型是基于将行为分解成由简单而直接的关系联系起来的事件。在复杂的社会技术系统中，这些事件链模型不足以理解、建模、评估和决策风险，因为这些系统中的事件具有间接和非线性的关系。需要更复杂的人类、组织和技术行为模型来理解和分析风险的组成部分，并提供在面对持续的系统适应和变化时做出良好决策所需的信息。本研究将探讨在面对持续变化和适应时，使用系统理论模型而不是分解模型来整合社会、组织、心理和工程方法来建模、分析和决策风险的潜力。我们将通过使用两个真实系统：NASA航天飞机计划和空中交通管制来验证这种方法，以提高对事故中涉及的复杂社会技术相互作用的理解。此外，我们将开发和实验评估基于模型的原型工具，以提高复杂系统的风险分析和决策制定能力。系统安全的系统理论方法为复杂系统中人的风险管理和决策的新一代研究奠定了基础。将社会、组织和技术因素整合到事故因果模型中，可以为人类和社会动力学研究开辟新的风险建模、分析和评估工具的大门。

Traditional models of causality are based on decomposition of behavior into events connected by simple and direct relationships. These chain-of-events models are inadequate to understand, model, assess, and make decisions about risk in complex socio-technical systems where the events have indirect and non-linear relationships. More sophisticated models of human, organizational, and technological behavior are needed to understand and analyze the components of risk and to provide the information required to make good decisions in the face of continual system adaptation and change. This research will examine the potential to use systems-theoretic rather than decompositional models to integrate social, organizational, psychological, and engineering approaches to modeling, analysis, and decision-making about risk in the face of continual change and adaptation. We will validate this approach to improved understanding of the complex socio-technical interactions involved in accidents by using two real systems: the NASA Space Shuttle Program and air traffic control. In addition, we will develop and experimentally evaluate prototype tools based on the model for their ability to improve risk analysis and decision making for complex systems. The systems-theoretic approach to system safety sets the groundwork for a new generation of research on human risk management and decision making in complex systems. The integration of social, organizational, and technical factors into a combined accident causality model could open the door to the development of new risk modeling, analysis, and assessment tools for human and social dynamics research.