Optimal resource allocation and technology adoption in dynamic systems

我的研究目标是在动态系统的背景下开发一系列决策框架，特别是由人口趋势、行为变化或技术进步驱动的系统。在这个研究项目中，我将开发运筹学模型，以解决动态决策的三个具有挑战性的特征：1)动态系统中信息分析的价值；2)结合技术变化预期的技术采用模型；以及3)多决策者框架，以发展对零散决策系统中技术采用的理解。我将开发的运筹学模型和分析方法来研究这些管理问题，很容易适应和推广到公共和私营部门问题的响应性决策。 这些研究中的每一个都受到现实世界决策问题的启发。因此，除了方法上的进步，我的工作还有可能对政策产生重大影响。 动态系统中信息的价值：这项工作的灵感来自于何时对传染病进行流行病学监测以便主动应对“暴发”这一难题。确定最佳预防和治疗策略的传统方法依赖于忽视随机过程的全部后果的确定性模型。在方法上，这一工作流将整合贝叶斯方法来评估信息价值，使用随机非线性动态系统进行传染病建模，以及最优控制理论。 技术采用与技术变革的预期：我通过调查储备使用技术替代品的存在如何影响最优技术消费决策，通过管理使用影响技术的长期价值的能力，以及通过研究和开发补贴影响新技术“到来”的速度来处理这个问题。在许多领域都遇到了这一问题，包括自然资源管理、信息技术基础设施以及全球抗菌素耐药性危机。 零散决策系统中的技术采用：保险覆盖的预防性服务的提供零散决策可能导致对具有长期效益的技术的次优投资，因为决策者既不承担所有的长期后果，也不能获得投资的所有好处。通过决策者之间的合同，特别是第二个保险提供商向第一个保险提供商的付款，可以纠正系统碎片化造成的低效。在这一系列研究中，我将这一框架扩展到包括可能影响合同克服零散决策能力的现实因素，如与技术制造商谈判价格的能力差异和信息不对称。

The goal of my research is to develop a series of frameworks for decision making in the context of dynamic systems, specifically systems driven by population trends, behavior change, or technological advancement. In this program of research, I will develop operations research models to address three challenging features of dynamic decision making: 1) value of information analysis in dynamic systems; 2) technology adoption models incorporating the expectation of technological change; and 3) multiple-decision-maker frameworks to develop an understanding of technology adoption in fragmented decision systems. The operations research models and analytical methods I will develop to investigate these management questions are readily adapted and generalized to responsive decision making across public and private sector problems. Each of these streams of research is inspired by a real-world decision problem. As a result, in addition to methodologic advancements, my work has the potential to have substantial policy impact. Value of information in dynamic systems: This stream of work is inspired by the difficult problem of when to perform epidemiologic monitoring of an infectious disease in order to proactively respond to an “outbreak”. Traditional approaches to identifying optimal prevention and treatment strategies rely on deterministic models overlooking the full consequences of the stochastic process. Methodologically, this stream of work will integrate Bayesian approaches to value-of-information, infectious disease modeling using stochastic non-linear dynamic systems, and optimal control theory. Technology adoption with the expectation of technological change: I approach this problem by investigating how optimal technology-consumption decisions are influenced by the existence of reserve-use technology alternatives, the ability to influence the long-term value of a technology through managed use, and the ability to influence the rate new technologies “arrive” through research and development subsidies. This problem is encountered in many areas, including management of natural resources, information technology infrastructure, and of the global crisis of antimicrobial resistance. Technology adoption in fragmented decision systems: Fragmented decision making in the provision of preventive services covered by insurance may cause suboptimal investment in technologies with long-term benefit because neither decision maker bears all the long-term consequences nor obtains all the benefits of investment. Correcting for the inefficiency caused by system fragmentation may be possible with contracts between decision makers, specifically a payment from the second insurance provider to the first. In this stream of research, I extend this framework to include realistic elements that may affect the ability of contracts to overcome fragmented decision making such as differences in ability to negotiate prices with technology manufacturers and information asymmetry.