||Distributed model predictive control (DMPC) is an emerging methodology that aims at designing localized controllers for a large-scale system, which in general is composed of constrained, multivariable and possibly even nonlinear subsystems that have a global objective and coupling constraints defined among them. Such systems can be found in areas as diverse as manufacturing equipments in the process industry, multiple vehicles in observation and monitoring systems such as mobile sensor networks, and smart structures in civil engineering. A key idea in DMPC schemes is to reduce the computational complexity associated with a centralized solution by solving local predictive control problems of small size based on models of subsystems and their neighbors.
This MSc project may consist of studying the achievable performace bounds by DMPC schemes and the trade-off between performance versus uncertainty among neighboring subsystems. The MSc project involves a literature study on decomposition methods in optimization and dynamic programming, which could be employed in a DMPC framework for an application of choice.