Collaborative Research: CNS Core: Small: Dynamic Pricing and Procurement for Distributed Networked Platforms

合作研究：CNS 核心：小型：分布式网络平台的动态定价和采购

基本信息

批准号：
2102963
负责人：
Mohammadhassan Hajiesmaili
金额：
$ 20万
依托单位：
University of Massachusetts Amherst
依托单位国家：
美国
项目类别：
Standard Grant
财政年份：
2021
资助国家：
美国
起止时间：
2021-10-01 至 2024-09-30
项目状态：
已结题

来源：
https://www.nsf.gov/awardsearch/showAward?AWD_ID=2102963&HistoricalAwards=false
关键词：
Collaborative Research CNS Core Small

项目摘要

Many industries today feature some kind of networked platform, where consumers may purchase resources from a network of providers. For example, mobile edge users can form a network and rent their compute resources to consumers. In order for these emerging businesses to survive and grow, however, they should ensure that their distributed resources are priced properly and made available in the proper amounts to users. Otherwise, some providers may find themselves overloaded with users and unable to serve their demands. This collaborative project between Carnegie Mellon University (CMU) and the University of Massachusetts Amherst (UMass) seeks to design foundational pricing, procurement, and scheduling policies that ensure that users are distributed well across the platform and apply these policies to the emerging application of edge computing.Optimal dynamic pricing schemes can signal to users which providers have resources available, while conversely dynamic procurement allows networked providers to adjust their resources to user demands. Scheduling schemes complement pricing and procurement solutions by leveraging time flexibility in user demands to best allocate resources to users. While several works have separately considered optimal pricing and scheduling policies for networked platforms, this project is the first to develop foundational theories for the joint formulation of dynamic pricing/procurement and scheduling under uncertainty. This project will develop pricing, procurement, and scheduling algorithms with theoretical performance guarantees; combine these solutions with learning-based approaches to manage tradeoffs between robustness and performance; and validate these solutions in edge computing scenarios.Successful development of the proposed pricing, procurement, and scheduling solutions will make the business of edge computing more profitable and competitive. Providers may gain insights into how to best price their resources, while users may gain flexibility that helps lower the cost of fulfilling their demands. Further, the theoretical tools developed will make foundational contributions to online optimization and learning research. In addition to these technical broader impacts, the project will support several education and outreach activities. These will include undergraduate research projects, integration of the research findings into courses at the participating institutions, and presentations and interactive sessions at workshops aimed at broadening participation in computing.The results of this project will be maintained in an online repository to be hosted by either UMass or CMU. These are expected to include technical reports of the research findings, software prototypes of the algorithms designed, and datasets and experimental results collected for the edge computing experiments. The material in the repository will remain available for at least two years after the project concludes.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.

当今许多行业都具有某种网络平台，消费者可以在其中从提供商网络购买资源。例如，移动边缘用户可以组建网络并将其计算资源出租给消费者。然而，为了让这些新兴企业生存和发展，他们应该确保其分布式资源的定价合理，并以适当的数量提供给用户。否则，一些提供商可能会发现自己的用户过多，无法满足他们的需求。卡内基梅隆大学 (CMU) 和马萨诸塞大学阿默斯特分校 (UMass) 之间的合作项目旨在设计基础定价、采购和调度政策，确保用户在整个平台上良好分布，并将这些政策应用于边缘计算的新兴应用。最佳动态定价方案可以向用户发出哪些提供商拥有可用资源的信号，而相反，动态采购允许网络提供商根据用户需求调整其资源。调度方案通过利用用户需求的时间灵活性来为用户最佳地分配资源，从而补充定价和采购解决方案。虽然有几项工作分别考虑了网络平台的最优定价和调度策略，但该项目是第一个开发不确定性下联合制定动态定价/采购和调度的基础理论的项目。该项目将开发具有理论性能保证的定价、采购和调度算法；将这些解决方案与基于学习的方法相结合，以管理鲁棒性和性能之间的权衡；并在边缘计算场景中验证这些解决方案。所提出的定价、采购和调度解决方案的成功开发将使边缘计算业务更具盈利能力和竞争力。提供商可以深入了解如何对其资源进行最佳定价，而用户可以获得有助于降低满足其需求的成本的灵活性。此外，开发的理论工具将为在线优化和学习研究做出基础性贡献。除了这些更广泛的技术影响外，该项目还将支持多项教育和外展活动。这些将包括本科生研究项目、将研究成果整合到参与机构的课程中，以及旨在扩大计算参与的研讨会上的演示和互动会议。该项目的结果将保存在由以下机构主办的在线存储库中：麻省大学或卡耐基梅隆大学。预计这些内容包括研究结果的技术报告、设计的算法的软件原型以及为边缘计算实验收集的数据集和实验结果。项目结束后，存储库中的材料将保留至少两年。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。

项目成果

期刊论文数量（15）

专著数量（0）

科研奖励数量（0）

会议论文数量（0）

专利数量（0）

Applied Online Algorithms with Heterogeneous Predictors

具有异构预测器的应用在线算法

DOI：
发表时间：
2023-07
期刊：
Proceedings of the 40th International Conference on Machine Learning
影响因子：
0
作者：
Jessica Maghakian; Russell Lee
通讯作者：
Russell Lee

Cooperative Stochastic Bandits with Asynchronous Agents and Constrained Feedback

具有异步代理和约束反馈的合作随机强盗

DOI：
发表时间：
2021-12
期刊：
Advances in Neural Information Processing Systems 34 (NeurIPS 2021
影响因子：
0
作者：
Lin Yang; Yu
通讯作者：
Yu

The War of the Efficiencies: Understanding the Tension between Carbon and Energy Optimization

效率之战：了解碳与能源优化之间的紧张关系

DOI：
10.1145/3604930.3605709
发表时间：
2023-07
期刊：
HotCarbon '23: Proceedings of the 2nd Workshop on Sustainable Computer Systems
影响因子：
0
作者：
Hanafy, Walid A.;Bostandoost, Roozbeh;Bashir, Noman;Irwin, David;Hajiesmaili, Mohammad;Shenoy, Prashant
通讯作者：
Shenoy, Prashant

Hierarchical Learning Algorithms for Multi-scale Expert Problems

多尺度专家问题的分层学习算法