Robust and Optimal Resource Allocation over Networks Subject to Externalities
Abstract
Robust allocation of scarce resources subject to network externalities constitutes a significant chalÂlenge in engineering and many other related fields. Robustness ensures that the distribution of resources is such that failure at some agents does not significantly affect the performance of the entire multiagent network system. Externalities capture the effect that networked agents may have on each other as a result of different resource allocation strategies. Externalities exist in many netÂwork systems, such as social, economic, and cyber-physical networks, and can substantially affect resource allocation strategies and outcomes. The scope of the proposed research is to study robust and optimal resource allocation strategies subject to externalities under three major thrusts- i) roÂbust data placement in peer-to-peer networks; ii) welfare optimization under network externalities; and iii) resilience of cyber-physical network systems. Specific goals of this research agenda are- i) designing effective and robust resource allocation strategies that can cope with network externalÂities; ii) analyzing network resource allocation strategies in a strategic and dynamic environment; iii) devising polynomial-time algorithms for computing or approximating optimal resource allocaÂtion strategies, and iv) validating the efficiency of the devised algorithms on real-world networks. Despite remarkable progress in the analysis of network resource allocation, the existing results mainly focus on simplified environments such as static networks, homogeneous agents, or netÂworks without externalities. In particular, the existing simplified algorithms often have poor perÂformance when applied to dynamic environments in which the network structure and allocation strategies are highly correlated. This research promises to provide the necessary mathematical foundations to extend the existing results on network resource allocation from the static homogeÂneous setting to dynamic heterogeneous settings subject to externalities. The proposed work will leverage tools from control theory, optimization theory, and network science to design robust and efficient resource allocation strategies over networks that exhibit externalities. This work will have direct impact on specific problems such as analyzing cascade failure in power grids and bandwidth allocation in autonomous vehicular networks interacting in new terrains of interest to the AF, or analyzing social network-transcendent behavioral dynamics of interest to the DoD. The developed tools will have a broader impact on many other related problems, such as control of fake news in social networks, security of power networks, facility location in distribution networks, and marÂketing in socioeconomic networks.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Feb 29, 2024
- Source ID
- FA95502310107
Entities
People
- S. Rasoul Etesami
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- University of Illinois Urbana–Champaign