EAGER: Improving Resource Utilization Through Peer-to-Peer Resource Sharing Systems

Rensselaer Polytechnic Institute, Troy NY

Investigators

Abstract

A novel way to increase resource efficiency is through the use of existing, idle capacity. A Peer-to-Peer Resource Sharing System is one where a resource owned by an individual is collectively shared with a group of users. The shared resource can be a physical resource (like a power drill) or a human resource (like the ability to perform a task). By focusing on access over ownership, these Systems allow physical assets to be consumed as services and tasks to be completed by independent individuals. A fundamental characteristic in these Systems is that both supply and demand originate from independent freelance users. Therefore, to match supply with demand in these Systems, a sequential process is common. A central mechanism recommends a set of available alternatives to decentralized users, who then select an alternative (or none at all) based on the user's valuation of the alternatives. This EArly-concept Grant for Exploratory Research (EAGER) award supports fundamental research to understand how to effectively match independent individuals through a central mechanism that allows for user selection. The new knowledge gained through this research can lead to Peer-to-Peer Resource Sharing Systems that are better understood, designed, and operated, which has the potential to positively impact important systemic societal issues such as transportation congestion, rising population increases, and limited natural resources. As the younger generation is early adopters of these systems, the integration of this research with educational and outreach activities has the potential to excite a new generation of engineers. The research will improve understanding of how to model and incorporate independent freelance user behavior into an optimization framework and introduce a new class of problems that considers how a central mechanism recommends a set of alternatives to decentralized users for selection. It will develop high fidelity models that incorporate theories and methods from the fields of recommender systems and consumer choice models to describe user preferences and selection behaviors in Peer-to-Peer Resource Sharing Systems. An outcome of decentralized user selection is that a 1-to-1 match of supply to demand does not always occur. Thus, a new understanding on how to determine the best way to rectify the demand requests and supply slots that have either been rejected or have duplicate selections will be explored. In doing so, a number of questions about the design of sequential, simultaneous, and hybrid protocols that match decentralized supply with demand will be answered.