CAREER: Systemic Risk and Strategic Formation in Stochastic Networks

该学院早期职业发展（CAREER）奖将通过研究复杂网络系统的有效运作，为国家繁荣和经济福利的发展做出贡献。当前的供应链网络、金融网络和其他行业具有复杂的互联结构，其中一个参与者的表现可能影响整个系统的表现。 在这种情况下，当一个供应商或金融机构的倒闭波及整个网络时，就会产生系统性风险，这可能会对整体福利造成代价高昂的影响。该奖项支持对管理网络漏洞，减轻系统性风险和解决故障的激励措施的基本理解，同时允许个人参与者的利润最大化行为。这项研究的跨学科性质将为学者、从业人员和政策制定者之间的沟通创造新的渠道，从而在设计有效政策时产生协同增效作用。伴随的教育计划旨在通过基于实验室课程和数字图书馆的实践活动来扩大STEM对随机网络的兴趣，并为代表性不足的社区提供机会。这项研究将建立一个包含广泛网络类别的战略决策框架，并将开发技术，为系统性风险缓解提供及时的解决方案。该框架将模拟网络中代理的反应，因为他们的联合决策过程适应冲击和故障，并考虑代理面临的网络状态的不完整信息。 拟议的研究填补了网络文献中的一个重要空白，主要考虑代理，既机械地遵循预先指定的协议，透明地观察整个网络。 分析基础设施利用并扩展了来自随机分析、博弈论和风险管理的最先进技术。该项目将开发数学技术，通过非线性不连续函数定点方程的解，在约束、加权和定向随机网络中恢复纳什均衡。它将产生数值算法来近似当代理不确定网络状态时出现的序列均衡，并根据其他代理及其相互作用的信念更新其高阶信念。该奖项反映了NSF的法定使命，并被认为是值得通过使用基金会的知识价值和更广泛的影响审查标准进行评估的支持。

This Faculty Early Career Development (CAREER) award will contribute to the advancement of national prosperity and economic welfare by studying efficient operations of complex network systems. Current supply chain networks, financial networks and other industries have complex interconnected structures, where the performance of one participant can affect the performance of the entire system. In such cases, systemic risk arises when the failure of one supplier or financial institution cascades through the network, which can be costly to overall welfare. This award supports a fundamental understanding of incentives that manage network vulnerabilities, mitigate systemic risk, and resolve failures, while allowing for profit-maximizing behavior of individual participants. The interdisciplinary nature of this research will create new channels of communications between academics, practitioners, and policy makers, leading to synergistic efficiencies in the design of effective policies. The accompanying educational plan aims to broaden STEM interest in stochastic networks through hands-on activities based on lab curricula and digital libraries, and to provide opportunities for underrepresented communities.This research will build a strategic decision-making framework encompassing a broad class of networks, and will develop techniques to provide timely solutions for systemic risk mitigation. The framework will model the reaction of agents in the network as their joint decision-making process adapts to shocks and failures, and account for agents' facing incomplete information on the state of the network. The proposed research fills an important gap in the network literature, which mostly consider agents that both mechanically follow pre-specified agreements and transparently observe the entire network. The analytical infrastructure leverages and extends state-of-the-art techniques from stochastic analysis, game theory, and risk management. The project will develop mathematical techniques to recover Nash equilibria in constrained, weighted, and directed stochastic networks through the solution of non-linear discontinuous functional fixed-point equations. It will produce numerical algorithms to approximate the sequential equilibria that arise when agents are uncertain about the state of the network, and update their higher-order beliefs on beliefs of other agents and their interactions. The performance assessments of the control policies will be informed by available data.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.