|
Wednesday, 24 March 2010 |
|
|
Colloquium:
|
Decomposition methods for multi-agent model predictive control
|
|
|
|
Speaker:
|
Dang Doan
|
|
|
|
Location:
|
room L
|
|
|
|
Time:
|
12:30 until 13:30
|
|
|
|
Abstract:
|
Model predictive control (MPC) is an emerging control technology which is able to deal with hard constraints, thus it is widely used in practice. The essence of MPC is to solve an online optimization problem at each sampling interval, in order to find the best control actions within a near future horizon. When using MPC in a multi-agent setup, i.e. there are multiple subsystems and each of them has a local control unit, one needs to use a decomposition method to let local controllers solve the centralized optimization problem. In this presentation, we provide a summary on state-of-the-art decomposition methods that are suitable for multi-agent MPC, and open topics for future research.
|
|
