A comparative analysis of distributed MPC techniques applied to the HD-MPC four-tank benchmark

作者：Alvarado I; Limon D^*; Munoz de la Pena D; Maestre J M; Ridao M A; Scheu H; Marquardt W; Negenborn R R; De Schutter B; Valencia F; Espinosa J

来源：Journal of Process Control, 2011, 21(5): 800-815.

DOI：10.1016/j.jprocont.2011.03.003

• Distributed control
• Predictive control
• Optimal control
• Benchmark examples
• Control applications

Recently, there has been a renewed interest in the development of distributed model predictive control (MPC) techniques capable of inheriting the properties of centralized predictive controllers, such as constraint satisfaction, optimal control, closed-loop stability, etc. The objective of this paper is to design and implement in a four-tank process several distributed control algorithms that are under investigation in the research groups of the authors within the European project HD-MPC. The tested controllers are centralized and decentralized model predictive controllers schemes for tracking and several distributed MPC schemes based on (i) cooperative game theory, (ii) sensivity-based coordination mechanisms, (iii) bargaining game theory, and (iv) serial decomposition of the centralized problem. In order to analyze the controllers, a control test is proposed and a number of performance indices are defined. The experimental results of the benchmark provide an overview of the performance and the properties of several state-of-the-art distributed predictive controllers.

• 出版日期2011-6