CAREER: An Adaptive Stochastic Look-ahead Framework for Disaster Relief Logistics under Forecast Uncertainty

该教师早期职业发展计划（CAREER）赠款将通过为飓风和缓慢移动的风暴等预先通知的自然灾害提供有效的救灾后勤行动的新知识，为国家健康，繁荣和福利的进步做出贡献。改进救灾工作既能减轻人类痛苦，又能减少经济损失。目前的救灾后勤规划和行动没有有效地纳入不断变化的天气预报和自然灾害分析工具。该项目将通过创建自适应决策支持方法来解决这一缺点，以有效地分期和利用稀缺资源，利用实时预测信息和历史数据。 该项目将通过设计新颖的物流决策支持工具，促进运筹学界和应急管理机构之间的长期合作。伴随的教育计划旨在通过数据驱动的分析工具丰富工程课程，创造跨学科研究机会，并为K开展外联活动-12名学生和公众，以帮助他们了解运筹学在解决关键的社会挑战，如救灾物流的作用。这项研究将有助于一个整体的建模和算法框架，顺序决策，在不断变化的灾害情况下的救灾后勤规划和行动及其滚动预测。该项目将：（i）建立新的理论来理解不断变化的预测不确定性对由过去的预测信息引起的决策政策质量的影响;（ii）产生新的算法，该算法在滚动时域过程中使用自适应采样、状态空间近似和阶段近似来集成离线和在线随机规划模型;以及（iii）创建和分析新颖的结构化决策策略，以解决协调具有异构模式的各种物流操作的时间安排的需要。将使用过去飓风的历史数据和模拟数据来验证救灾物流业务规划的建模和解决方法。研究结果将有助于应急管理人员参与和制定物流规划和运营政策，在实践中平衡适应性，最优性和可执行性。该奖项反映了NSF的法定使命，并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，被认为值得支持。

This Faculty Early Career Development Program (CAREER) grant will contribute to the advancement of national health, prosperity and welfare by contributing new knowledge on effective disaster relief logistics operations for advance-notice natural disasters such as hurricanes and slow-moving storms. Improved disaster relief efforts can both alleviate human suffering and reduce economic loss. Current disaster relief logistics planning and operations do not effectively incorporate evolving weather forecasts and natural hazard analysis tools. This project will address this shortcoming by creating adaptive decision-support methods for effectively staging and utilizing scarce resources, leveraging both real-time forecast information and historical data. This project will foster a long-term collaboration between the operations research community and emergency management agencies by designing novel logistics decision support tools. The accompanying educational program aims to enrich engineering curriculum with data-driven analytic tools, create interdisciplinary research opportunities, and develop outreach activities for K-12 students and the general public to help them understand the role of operations research in addressing critical societal challenges such as disaster relief logistics.This research will contribute a holistic modeling and algorithmic framework for sequential decision making in disaster relief logistics planning and operations under dynamically evolving disaster situations and their rolling forecasts. This project will: (i) establish new theory to understand the impact of evolving forecast uncertainty on the quality of the decision policy induced by past forecast information; (ii) produce novel algorithms that integrate offline and online stochastic programming models using adaptive sampling, state space approximation, and stage approximation within a rolling-horizon procedure; and (iii) create and analyze novel structured decision policies to address the need to coordinate the timing of various logistics operations with heterogeneous modalities. The modeling and solution methodology on disaster relief logistics operations planning will be validated using both historical data on past hurricanes and simulation data. Research results will help engage and inform emergency managers in making logistics planning and operational policies that balance between adaptability, optimality and executability in practice.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.