Abstract: Fair allocation of resources is a central problem at the interface of economics and computer science. Two commonly studied fairness criteria are envy-freeness and equitability; the former requires that no agent prefers another agent's allocation to its own, while the latter strives to balance the absolute utilities of the agents. In addition to fairness, one might also desire economic efficiency, and the gold standard here is Pareto optimality. In this talk, we will discuss the allocation of indivisible (or discrete) items, and present an algorithmic framework for simultaneously achieving the seemingly incompatible notions of fairness and efficiency.
Based on joint works with Siddharth Barman and Sanath Kumar Krishnamurthy, and with Rupert Freeman, Sujoy Sikdar, and Lirong Xia.
Bio: Rohit Vaish is a postdoctoral researcher at Rensselaer Polytechnic Institute. Prior to that, he did his PhD from Indian Institute of Science. His research interests include problems in economics and computation, specifically in the area of computational social choice.