NYC Gazette

Researchers have introduced a game theory model aimed at optimizing reward policies in the sharing economy. The team, led by Dr. David Li of the Katz School's M.S. in Data Analytics and Visualization, presented their work at the 59th Annual Conference on Information Science and Systems at Johns Hopkins University. The study, “A Dynamic Framework for Optimizing Reward Policies in the Sharing Economy,” involved contributions from Chengkun Yao and Cheng Li, both master's students at Katz School, and Angela Li from Stony Brook University.

The sharing economy allows for temporary access to services through platforms that depend heavily on the engagement of service providers, which is driven by incentives. In response to this, Dr. Li said, "Static reward models, where incentives remain fixed regardless of market conditions, often fail to adapt to evolving user behaviors and demand fluctuations. Our study addresses the limitations of such models by introducing a dynamic framework that adjusts rewards in real-time, maximizing engagement while keeping costs under control."

This model incorporates game theory principles, particularly Nash equilibrium, to predict user responses to varying incentives, suggesting a shift from static to adaptive policies. In contexts such as grocery delivery services, this means dynamically increasing rewards during worker shortages to ensure more workers, and reducing them during oversupply to control costs.

Chengkun Yao noted, "To implement this adaptive model, we formulated the reward allocation problem as an optimization challenge. Using a dynamic programming algorithm, the framework recursively selects which users to incentivize, aiming to maximize platform efficiency and profitability while adhering to budget constraints."

Historical data and machine learning aid in optimizing incentives, helping balance demand and supply. Cheng Li highlighted, "The results demonstrated diminishing returns at high participation levels, indicating that beyond a certain point, additional incentives fail to yield proportional increases in engagement." This insight calls for targeting key users contributing significantly at lower costs.

The model's adaptability is particularly beneficial for smaller platforms under financial constraints, ensuring optimal engagement through strategic resource allocation. Angela Li commented, "By transitioning from static to dynamic reward policies, platforms can unlock new levels of efficiency, profitability, and user satisfaction, ensuring sustainable growth in an increasingly competitive market."

The framework promises benefits including maximized engagement, cost efficiency, scalability, and fairness. It supports seamless adjustments and equitable reward distribution, fostering trust and retention among platform users, potentially reshaping engagement strategies within the sharing economy.