Multi-agent Predictive Control with Application in Intelligent Infrastructures
This paper describes a stabilizing decentralized predictive controller for multivariable systems with application in intelligent infrastructures. Cooperative distributed control systems can be seen as a set of computational agents that cooperate with thei
- PDF / 227,739 Bytes
- 10 Pages / 439.37 x 666.142 pts Page_size
- 91 Downloads / 183 Views
Multi-agent Predictive Control with Application in Intelligent Infrastructures J.M. Igreja, S.J. Costa, J.M. Lemos, and F.M. Cadete
Abstract This paper describes a stabilizing decentralized predictive controller for multivariable systems with application in intelligent infrastructures. Cooperative distributed control systems can be seen as a set of computational agents that cooperate with their neighbors, in a open information interchange infrastructure, to achieve local performance indexes, suitable to globally control large scale, geographically expansive, systems. An agent behavioral induction approach is introduced and discussed in the model predictive control framework. Cooperative iteration convergence, with information interchange, is proved. The algorithm is fully developed for a serially chained systems and an application to a real water delivery canal is presented.
12.1
Introduction
This paper describes a stabilizing decentralized predictive control for multivariable systems with application to intelligent infrastructures. Cooperative distributed control systems can be seen as a set of computational agents that cooperate with their neighbors, in an open information interchange infrastructure, to achieve local performance indexes, suitable to globally control in closed-loop large scale, geographically expansive, systems. A behavioral induction approach is introduced and discussed in the model predictive control framework. The predictive control algorithm is fully developed for generic serially chained systems. J.M. Igreja (*) Automation and Power Systems Engineering Department, ISEL, R. Conselheiro Emdio Navarro, 1, Lisbon, Portugal e-mail: [email protected] S.J. Costa Chemical Engineering Department, ISEL, R. Conselheiro Emdio Navarro, 1, Lisbon, Portugal J.M. Lemos • F.M. Cadete Control of Dynamic Systems Group, INESC-ID/IST, R. Alves Redol, 9, Lisbon, Portugal e-mail: [email protected] A. Madureira et al. (eds.), Computational Intelligence and Decision Making: Trends and 121 Applications, Intelligent Systems, Control and Automation: Science and Engineering 61, DOI 10.1007/978-94-007-4722-7_12, # Springer Science+Business Media Dordrecht 2013
122
J.M. Igreja et al.
Decentralized control systems arise from the need to deal with a large number of interconnected geographical distributed subsystems where information exchange between local subsystems is a priori somehow constrained. This physical inability decentralized framework since each system is often simpler to understand than a large composite MIMO system. Fault tolerance, growing geographical expansive systems, large scale networked systems, such as power systems grids and water distribution canal pools, are other good reasons to decentralize control. David Fogel’s book [3] dedicates one chapter to closed-loop control and model prediction recognizing that “Intelligence requires iterative prediction and control” as two basic features. In this paper computational agents include, beside these two, cooperative behavior for obtaining the same com
Data Loading...
