PFI-RP: A Smart Food Distribution System for Allocating Scarce Resources Under Extreme Events

创新伙伴关系-研究伙伴关系（PFI-RP）项目的更广泛影响/商业潜力是为从事提供粮食援助的组织提供直接利益，这将为人民和社会带来更好的结果，特别是在极端事件期间增加弱势群体获得粮食的机会方面。尽管美国200多家食品银行在2020年救助了近40亿磅的食品，但食品不安全仍是一个长期存在的问题。通过利用食物银行的实际数据、对食物银行运作的了解以及人工智能（AI）的使用，开发的系统将提供有关食物收集、储存和分配的政策建议，这些建议将在食物分配中体现公平，同时降低获取和分配食物的成本，并减少食物浪费。这些建议将改变组织如何根据环境破坏快速调整其运营。拟议的项目将创建并增强智能食品分配系统的弹性，以便在极端事件下直接根据数据进行资源分配。将开发一个通用框架，用于聚合和标准化跨各种媒体（操作数据，公开可用）的数据收集，以提供（大规模）供应和需求的自适应预测。开发的数据驱动优化模型将促进食品分配策略的近实时评估，并提供食品分配网络的材料、财务和物理能力的交互式和可解释的模型可视化。负责大中型饥饿救济组织供应链的管理人员将受益于人工智能驱动的资源管理软件，以增加为粮食不安全家庭提供服务的数量。此外，该项目将为决策研究提供信息，最终使饥饿救济组织在极端情况下更具弹性，从而减轻供应链网络中断的负面影响。该奖项反映了美国国家科学基金会的法定使命，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。

The broader impact/commercial potential of this Partnerships for Innovation - Research Partnerships (PFI-RP) project is to provide direct benefit to organizations engaged in the provision of food aid, which will translate to better outcomes for people and society, particularly with respect to increasing access to food for vulnerable populations during extreme events. Food insecurity is a persistent problem in the US despite the nearly 4 billion pounds of food rescued in 2020 by over 200 food banks. By leveraging actual data from the foodbanks, knowledge of the operations of foodbanks and the use of Artificial Intelligence (AI), the developed system will provide policy recommendations about the collection, storage, and distribution of food that will incorporate fairness in food distribution while reducing the cost of acquiring and distributing food, and cutting down food waste. These recommendations will transform how organizations rapidly adapt their operations according to environmental disruptions. The proposed project will create and enhance the resilience of smart food distribution systems for resource allocation under extreme events directly from data. A generalizable framework for aggregating and standardizing the collection of data across various media (operational data, publicly available) to provide adaptive predictions of supply and need (at scale) will be developed. The developed data-driven optimization models will facilitate near real-time evaluation of food distribution strategies and provide interactive and interpretable model visualizations of the material, financial, and physical capacity of the food distribution networks. Managers responsible for supply chains in medium to large hunger relief organizations will benefit from the Al-driven resource management software to increase the number of food insecure households served. Additionally the project will inform research on decision-making which will ultimately make hunger relief organizations more resilient in extreme situations thus mitigating the negative impacts of disruptions to supply chain networks.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.