CRISP Type 1/Collaborative Research: Financial and Physical Infrastructure: A Computational Approach for Integrated Network Resilience Analysis Under Extreme Events

极端事件对社会的影响不仅是对有形基础设施系统造成的直接损害的作用，也是社区能够实施的灾后恢复进程的作用。这一复苏进程取决于金融部门(如私人银行、保险和再保险公司)提供的投资。因此，这些金融基础设施和有形基础设施之间的适当关系是必要的，以便为迅速和有效地恢复有形系统提供财务先决条件。另一方面，这些极端事件也是金融网络的压力源，如果金融基础设施和有形基础设施之间的关系没有以可持续、有弹性的方式建立起来，就可能导致崩溃。该项目旨在加深对金融基础设施和有形基础设施之间的关系及其对这些相互关联的系统的复原力的影响的理解。这一认识的提高最终将导致对决策者以及金融和有形基础设施管理人员的支持，以便改进对极端事件的准备和反应能力。本项目的目标将通过将有形基础设施风险评估和金融网络建模领域的现有研究与大型互联系统分析的计算技术相结合来实现。极端事件下系统的工程可靠性分析模型和方法，系统演化和重大中断后恢复的马尔可夫链模型，以及运输和供应基础设施的网络模型将用于描述物理基础设施。将使用金融网络分析技术模拟金融基础设施以及这些机构之间以及实体和金融部门之间的合同关系，纳入最近在危机传播和逐步金融崩溃方面的结果。将这些整合到互连的物理和金融基础设施的通用异质网络模型中，将在一系列潜在危险情景下进行复原力分析。由于将在灾害恢复过程中对大型实物和金融资产系统的反应进行建模所涉及的计算挑战，将开发和校准使用图论方法分析的用于网络扩散分析的替代模型，从而能够有效地分析大型系统。最后，利用上述复原力分析，将对代表这些实体之间合同关系的实物-财务网络的图形结构进行优化，以最大限度地提高所产生的系统在极端危险事件下的复原力。这种网络优化的结果将有助于决策者确定哪些合同结构和政策最能支持和提高互联系统的复原力。总体而言，该项目将从结合工程学、金融学和网络理论的跨学科角度，更好地理解灾害、脆弱性和融资在极端事件影响社区的灾后恢复中的相互作用。在计算上，该框架将允许分析大型系统，包括随机(或不确定)的异质网络结构。

The effects of extreme events on society are not only a function of the immediate damage induced on the physical infrastructure system, but also of the post-event recovery process that a community is able to implement. This recovery process depends on the availability of investments from the financial sector (e.g. by private banks, insurance, and re-insurance companies). Appropriate relationships between these financial and physical infrastructures are therefore necessary to provide the financial pre-conditions for rapid and efficient restoration of the physical system. On the other hand, these extreme events also represent stressors for the financial network, and can cause collapses if the relationships between financial and physical infrastructures are not established in a sustainable, resilient manner. This project aims at developing a deeper understanding of the relationships coupling the financial and physical infrastructures and their effect on the resilience of these interconnected systems. This improved understanding will ultimately lead to support for policy makers and for financial and physical infrastructure managers in order to improve preparedness for and responsiveness to extreme events.The aims of this project will be accomplished through the integration of existing research in the fields of physical infrastructure risk assessment and financial network modeling with computational techniques for the analysis of large interconnected systems. Models and methods for engineering reliability analysis of systems subjected to extreme events, Markov chain models for system evolution and recovery following major disruption, and network models of transportation and supply infrastructures will be used for describing the physical infrastructures. Financial infrastructures and the contractual relationships among these institutions and between physical and financial sectors will be modeled using financial network analysis techniques, incorporating recent results in distress propagation and progressive financial collapse. Combining these into a common heterogeneous network model for interconnected physical and financial infrastructures, resilience analysis will be conducted under a suite of potential hazard scenarios. Because of the computational challenges associated will modeling the responses of large systems of physical and financial assets during the hazard recovery process, surrogate models for network diffusion analysis, analyzed using graph-theoretical approaches, will be developed and calibrated, allowing for efficient analysis of large-scale systems. Finally, making use of the above resilience analysis, the graphical structure of the physical-financial network, representing the contractual relationships between these entities, will be optimized in order to maximize the resilience of the resulting system under the extreme hazard event. The results of this network optimization will be useful to policy makers in determining which contractual structures and policies best support and improve the resilience of the interconnected system. Overall, this project will result in a better understanding of the interactions of hazard, vulnerability, and financing in the post-event recovery of communities exposed to extreme events from an interdisciplinary perspective combining engineering, finance, and network theory. Computationally, the framework will allow for analysis of large systems, including random (or uncertain) heterogeneous network structures.