Thesis Seminar

The presentation covers the main chapters of my dissertation and will be broadly related to online platforms. 

In the first part of my talk, I will be providing insights on the trade-offs between transparency and control. In particular, we will be looking at three platform designs and their worst case efficiency loss (Price of Anarchy) under a networked Cournot competition model. 

The second part of my talk focuses on a platform on the electricity grid, and includes a study on the impact of anticipation and when it leads to manipulation in these markets. We incorporate network constraints that mimic constraints faced under the electricity grid setting. We will show that demand management, e.g., demand response, have the potential to reduce the impact and incentives for price manipulation. We study this under a Stackelberg version of the networked Cournot competition model. 

I will conclude the talk with recent work on an algorithm for an online optimization problem with inventory constraints. The problem generalizes the classical one-way trading problem, and aims to mimic (selling) decision making, such as in the participation of ancillary services, or driving for ride-sharing platforms. Our algorithm (CR-Pursuit) provides near-optimal algorithms for the generalized one-way trading problem.