Loading Events

ESE Guest Seminar – “On Team Decision Problems with Nonclassical Information Structures”

February 25, 2025 at 1:00 PM - 2:00 PM
Details
Date: February 25, 2025
Time: 1:00 PM - 2:00 PM
Event Category: SeminarColloquium
  • Event Tags:, ,
  • Organizer
    Electrical and Systems Engineering
    Phone: 215-898-6823
    Venue
    Greenberg Lounge (Room 114), Skirkanich Hall 210 South 33rd Street
    Philadelphia
    PA 19104
    Google Map

    Team theory is a mathematical formalism for decentralized stochastic control problems in which a “team,” consisting of a number of members, cooperates to achieve a common objective. It was developed to provide a rigorous mathematical framework of cooperating members in which all members have the same objective yet different information. In static team problems, the information received by the team members is not affected by the decisions of other team members, while in dynamic team problems, the information of at least one team member is affected by the decisions of other team members. If there is a prescribed order in which team members make decisions, then such a problem is called a sequential team problem. The information structures in sequential team decision problems designate who knows what about the status of the team and are classified as classical, partially nested, and non-classical. In classical information structures, all team members receive the same information and have perfect recall. In partially nested information structures, some team members have a nonempty intersection of their information structures while they have perfect recall. Any information structure that is not classical or partially nested is called nonclassical. In this talk, I consider sequential dynamic team decision problems with nonclassical information structures. First, I will address the problem from the point of view of a “manager” who seeks to derive the optimal strategy for the team in a centralized process. I provide structural results that yield an information state for the team, which does not depend on the control strategy, and thus, it can lead to a dynamic programming decomposition where the optimization problem is over the space of the team’s decisions. I will then provide structural results for each team member that yield an information state that does not depend on their control strategy, and thus, it can lead to a dynamic programming decomposition where the optimization problem for each team member is over the space of their decisions. Finally, I will show that the solution of each team member is the same as the one derived by the manager. Therefore, each team member can derive their optimal strategy, which is also optimal for the team, without the manager’s intervention.