Self-organization and Intervention of Nonlinear Multi-agent Systems
Ph.D. thesis defense pre-seminar: Yuecheng Yang will give a seminar where he presents his own view of his thesis before the defence.
Time: Fri 2016-11-18 11.00 - 12.00
Location: Seminar room 3721, Lindstedtsvägen 25
Participating: Yuecheng Yang
Abstract:
This talk concerns the self-organization behaviors in different types of multi-agent systems, and possible ways to apply interventions on top of that to achieve certain goals. In the first part of the presentation a bounded confidence opinion dynamics model is considered. Theoretical analysis of the model is performed and modifications of the model are given so that it will have better properties in some aspect. A possible intervention with an exo-system is also studied where sufficient conditions on the exo-system are given for the purpose of consensus. In the second part, optimal control problems with leader-follower based multi-agent systems are discussed. For general non-linear leader-follower interactions, we provide a method that uses statistic moments of the follower crowd to approximate the optimal control. The dynamic programming approach is used and certain approximation of the Hamilton-Jacobi-Bellman (HJB) equations is needed. The computational burden is so heavy that model predictive control method is required in practical applications. This approach can be used to attack optimal control problem with partial differential equation (PDE) constraints by discretization in space.