Research Topic Area 11.1: ARO SPECIAL PROGRAMS: STIR PROGRAM - Limited Learning, Rational Inattention, and Unawareness in Games and Decision Problems
Abstract
Game theory and decision theory have made major progress by assuming that all players are rational. Effectively this means that they are computationally unbounded, since the rationality assumption, as applied in the literature, means that players can compute their beliefs and a best response to what they believe other players are doing. Although this assumption is useful, it is well known that in many games and decision problems, it does not lead to correct predictions. To take just one well-known example, people do cooperate in the Prisoner s Dilemma. This proposal investigates models that take seriously the assumption that people are acting rationally, in the sense of doing the best that they can, subject to their computational limitations. The ultimate goal of this project is to provide a foundational theory for how people actually play games and make decisions in realistic settings, where resource limitations arise. Specific topics to be investigated include learning from limited windows (i.e., ignoring information that was not received in the recent past) and rational inattention (ignoring seemingly relevant attributes of a problem).
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 22, 2017
- Source ID
- W911NF1610397
Entities
People
- Joseph Halpern
Organizations
- Army Contracting Command
- Cornell University
- United States Army