Author ORCID Identifier

Date Available


Year of Publication


Degree Name

Doctor of Philosophy (PhD)

Document Type

Doctoral Dissertation




Computer Science

First Advisor

Dr. Simone Silvestri


Electric power systems are transforming from a centralized unidirectional market to a decentralized open market. With this shift, the end-users have the possibility to actively participate in local energy exchanges, with or without the involvement of the main grid. Rapidly reducing prices for Renewable Energy Technologies (RETs), supported by their ease of installation and operation, with the facilitation of Electric Vehicles (EV) and Smart Grid (SG) technologies to make bidirectional flow of energy possible, has contributed to this changing landscape in the distribution side of the traditional power grid.

Trading energy among users in a decentralized fashion has been referred to as Peer- to-Peer (P2P) Energy Trading, which has attracted significant attention from the research and industry communities in recent times. However, previous research has mostly focused on engineering aspects of P2P energy trading systems, often neglecting the central role of users in such systems. P2P trading mechanisms require active participation from users to decide factors such as selling prices, storing versus trading energy, and selection of energy sources among others. The complexity of these tasks, paired with the limited cognitive and time capabilities of human users, can result sub-optimal decisions or even abandonment of such systems if performance is not satisfactory. Therefore, it is of paramount importance for P2P energy trading systems to incorporate user behavioral modeling that captures users’ individual trading behaviors, preferences, and perceived utility in a realistic and accurate manner. Often, such user behavioral models are not known a priori in real-world settings, and therefore need to be learned online as the P2P system is operating.

In this thesis, we design novel algorithms for P2P energy trading. By exploiting a variety of statistical, algorithmic, machine learning, and behavioral economics tools, we propose solutions that are able to jointly optimize the system performance while taking into account and learning realistic model of user behavior. The results in this dissertation has been published in IEEE Transactions on Green Communications and Networking 2021, Proceedings of IEEE Global Communication Conference 2022, Proceedings of IEEE Conference on Pervasive Computing and Communications 2023 and ACM Transactions on Evolutionary Learning and Optimization 2023.

Digital Object Identifier (DOI)

Funding Information

National Institute for Food and Agriculture NIFA 2017-67008-26145

National Science Foundation grant EPCN 1936131

National Science Foundation CAREER grant CPS-1943035