Premium
Augmented Lagrangian Pattern Search Based Multi‐Agent Model Predictive Control of Rhine‐Meuse Delta
Author(s) -
Sharafi Narjes,
Safavi Ali Akbar
Publication year - 2016
Publication title -
asian journal of control
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.769
H-Index - 53
eISSN - 1934-6093
pISSN - 1561-8625
DOI - 10.1002/asjc.1327
Subject(s) - model predictive control , scheme (mathematics) , scale (ratio) , computer science , augmented lagrangian method , control (management) , nonlinear system , mathematical optimization , multi agent system , lagrangian , control system , control theory (sociology) , engineering , artificial intelligence , mathematics , algorithm , mathematical analysis , physics , quantum mechanics , electrical engineering , mathematical physics
In Large Scale Systems the concept of centrality fails due to the lack of centralized computing capability. The control of such systems has to be performed using multiple control agents. In this case, the matter of interactions among neighboring subsystems needs to be considered. In this paper, a water control system in the Netherlands is studied as a real large scale system. A multi‐agent scheme is applied to control the flow through the system which is decomposed into two interconnected subsystems. Each agent employs a model‐based predictive control (MPC) technique. The model of this large scale system is nonlinear and nonconvex. Therefore, an augmented Lagrangian pattern search optimization algorithm is used to implement multi‐agent MPC for this system. This proposed algorithm is applied by each control agent to solve its own interconnected optimization problem, at each subsystem of whole the water system. Simulation results show the effectiveness of the proposed approach.