Georgia Tech SCL Georgia Tech SCL Georgia Tech SCL

Title:

 Incentives for Exploration at Market Equilibrium.

Abstract:

This work attempts to understand the role of congestion in creating incentives for exploration amongst myopic customers in online marketplaces where the qualities of value-providing entities (e.g., sellers, service providers) are learned over their market lifetime. While equilibrium prices are generally effective in internalizing the congestion externalities resulting from the matchings, the extent to which they internalize the exploration externalities in such markets is unclear. We define a new market equilibrium notion called a local market equilibrium (LME) to investigate this question. After proving several existence and structural results for the LME, we study its performance. We first show market instances where the loss in matching reward under the LME can be unbounded relative to the system optimum. We then tightly characterize a market condition under which this relative loss is bounded below by the aggregate congestion level, i.e., the ratio of demand and aggregate supply in the market. We also show that it is impossible to improve upon this bound unless further structural assumptions are made. As examples, we define a broad class of market instances with linear information-state transitions and a monotonic reward structure in which the unique LME corresponds to system optimality. Finally, we address the problem of designing optimal targeted price interventions that align the LME with the system-optimal solution. Our results inform market designers grappling with the concern of incentivizing exploration in various online marketplaces.

Bio:

Vijay Kamble is an Assistant Professor in the Department of Information of Decision Sciences and a faculty affiliate in the Department of Computer Science at the University of Illinois Chicago. Before joining UIC in 2017, he was a postdoctoral scholar in the Society and Algorithms Lab at the Management Science and Engineering Department of Stanford University. He obtained his Ph.D. in Electrical Engineering and Computer Sciences from the University of California Berkeley in 2015. His research is centered on the design and optimization of online platforms and marketplaces, with a particular interest in characterizing algorithmic tradeoffs at the interfaces between incentives, learning, and the concerns of fairness and ethics in such platforms.