Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 21
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Contribución de congreso (ConferenceOutput)

Model predictive control of drinking water networks: A hierarchical and decentralized approach

  • Ocampo-Martínez, Carlos
  • Fambrini, Valentina
  • Barcelli, Davide
  • Puig, Vicenç
Trabajo presentado al ACC 2010 celebrado en Baltimore (EE.UU.) del 30 de junio al 2 de julio., In this paper, a decentralized model predictive control (DMPC) strategy for drinking water networks (DWN) is proposed. The DWN is partitioned in a set of subnetworks using a partitioning algorithm that makes use of the topology of the network, the information about the actuator usage and heuristics. A suboptimal DMPC strategy was derived that allows the hierarchical solution of the set of MPC controllers used to control each partition. A comparative study between the centralized MPC (CMPC) and DMPC approaches is developed on the case study, which consists in an aggregate version of the Barcelona DWN. Results have shown the effectiveness of the proposed DMPC approach in terms of the computation time while an admissible level of suboptimality is obtained in all the considered scenarios., This work was supported by projects: 'Contribució a la promoció i supervisió de projectes de recerca de l'àmbit de les xarxes de distribució d'aigua' (C-7006), 'Analysis and Design of distributed optimal control strategies applied to large-scale WATer systems MANagement' (4808). This work has been supported by the CICYT DPI 2009 -13744 of the Spanish Science Ministry, the Juan de la Cierva Research Programme (ref. JCI-2008-2438), the DGR of Generalitat de Catalunya (SAC group Ref. 2009/SGR/1491) and the EU project WIDE (FP7-IST-224168)., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Contribución de congreso (ConferenceOutput)

Distributed MPC for large scale systems using agent-based reinforcement learning

  • Javalera, Valeria
  • Morcego, Bernardo
  • Puig, Vicenç
Presentado al 12th IFAC Symposium on Large-Scale Systems: Theory and Applications celebrado en Francia en 2010., In the present work, distributed control and artificial intelligence are combined in a control architecture for Large Scale Systems (LSS). The aim of this architecture is to provide a general structure and methodology to perform optimal control in networked distributed environments where multiple dependencies between sub-systems are found. Often these dependencies or connections represent control variables so the distributed control has to be consistent for both subsystems and the optimal value of these variables has to accomplish a common goal. The aim of the research described in this paper is to exploit the attractive features of MPC (meaningful objective functions and constraints) in a distributed implementation combining learning techniques to perform the negotiation of these variables in a cooperative Multi Agent environment and over a Multi Agent platform to provide speed, scalability, and computational effort reduction. This approach is based on negotiation, cooperation and learning. Results of the application of this architecture to a small drinking water network show that the resulting trajectories of the levels in tanks (control variables) can be acceptable compared to the centralized solution. The application to a real network (the Barcelona case) is currently under development., This work was supported by the project 'Analysis and Design of distributed optimal control strategies applied to large-scale WATer systems MANagement' (4808). The authors wish to thank the support received by WIDE-224168-FP7-ICT-2007-2 project and WATMAN ref. DPI2009-13744 of Spanish Ministry of Education. First author is supported by CONACYT and partially supported by ITESCA., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Contribución de congreso (ConferenceOutput)

Decentralized model predictive control of drinking water networks using an automatic subsystem decomposition approach

  • Barcelli, Davide
  • Ocampo-Martínez, Carlos
  • Puig, Vicenç
  • Bemporad, Alberto
Presentado al 12th IFAC Symposium on Large-Scale Systems: Theory and Applications celebrado en Francia en 2010., This paper proposes an automatic model decomposition approach for decentralized model predictive control (DMPC) of drinking water networks (DWNs). For a given DWN, the proposed algorithm partitions the network in a set of subnetworks by taking advantage of the topology of the network, of the information about the use of actuators, and of system management heuristics. The derived suboptimal DMPC strategy results in a hierarchical solution with a set of MPC controllers used for each partition. A comparative study between centralized MPC (CMPC) and DMPC approaches is described for the considered case study, which consists of an aggregate version of the Barcelona DWN. Results on several simulation scenarios show the effectiveness of the proposed DMPC approach in terms of the reduced computation burden and, at the same time, of the admissible lost of performance., This work was supported by the project 'Analysis and Design of distributed optimal control strategies applied to large-scale WATer systems MANagement' (4808). This work has been partialy supported by the CICYT Ref. DPI2009-13744 of the Spanish Science and Technology Ministry, the Juan de la Cierva Research Programme (ref. JCI-2008-2438), the DGR of Generalitat de Catalunya (SAC group Ref. 2009/SGR/1491), and the European project WIDE, contract number FP7-IST-224168., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Contribución de congreso (ConferenceOutput)

Negotiation and Learning in distributed MPC of Large Scale Systems

  • Javalera, Valeria
  • Morcego, Bernardo
  • Puig, Vicenç
Congreso celebrado en Baltimore (USA) del 30 de Junio al 2 de Julio de 2010., A key issue in distributed MPC control of Large Scale Systems (LSS) is how shared variables among the different MPC controller in charge of controlling each system partition (subsystems) are handled. When these connections represent control variables, the distributed control has to be consistent for both subsystems and the optimal value of these variables will have to accomplish a common goal. In order to achieve this, the present work combines ideas from Distributed Artificial Intelligence (DAI), Reinforcement Learning (RL) and Model Predictive Control (MPC) in order to provide an approach based on negotiation, cooperation and learning techniques. Results of the application of this approach to a small drinking water network show that the resulting trajectories of the levels in tanks (control variables) can be acceptable compared to the centralized solution. The application to a real network (the Barcelona case) is currently under development., This work was supported in part by the WIDE - 224168-FP7-ICT-2007-2 project and by WATMAN ref. DPI2009-13744. Valeria is also supported by the Consejo Nacional de Ciencia y Tecnología (CoNACYT) México, D.F. and the Instituto Tecnológico Superior de Cajeme (ITESCA), Cd Obregón Son., México., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Contribución de congreso (ConferenceOutput)

A multi-agent MPC architecture for distributed large scale systems

  • Javalera, Valeria
  • Morcego, Bernardo
  • Puig, Vicenç
9 páginas, 7 figuras.-- Trabajo presentado al 2nd International Conference on Agents and Artificial Intelligence (ICAART) celebrado en Valencia (España) en Enero de 2010., In the present work, techniques of Model Predictive Control (MPC), Multi Agent Systems (MAS) and Reinforcement Learning (RL) are combined to develop a distributed control architecture for Large Scale Systems (LSS). This architecture is multi-agent based. The system to be controlled is divided in several partitions and there is an MPC Agent in charge of each partition. MPC Agents interact over a platform that allows them to be located physically apart. One of the main new concepts of this architecture is the Negotiator Agent. Negotiator Agents interact with MPC Agents which share control variables. These shared variables represent physical connections between partitions that should be preserved in order to respect the system structure. The case of study, in which the proposed architecture is being applied and tested, is a small drinking water network. The application to a real network (the Barcelona case) is currently under development., This work is in the context of the European Project Decentralized and Wireless Control of Large Scale Systems, WIDE--FP7-ICT-2007-2. First author is sponsored by the Consejo Nacional de Ciencia y Tecnología (CONACYT) of México and is partially supported by the Instituto Tecnológico Superior de Cajeme (ITESCA)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Contribución de congreso (ConferenceOutput)

Decentralised MPC based on a graph partitioning approach applied to the Barcelona drinking water network

  • Ocampo-Martínez, Carlos
  • Puig, Vicenç
  • Bovo, Samuele
Trabajo presentado al 18th IFAC World Congress celebrado en Milán del 28 de agosto al 2 de septiembre de 2011., In this paper, the application of decentralised model predictive control (DMPC) over a partitioned model of the Barcelona drinking water network (DWN) is proposed. The system decomposition is based on graph partitioning theory, which considers the graph representation of the DWN topology. Once obtained, this system decomposition yields to some bidirectional flows of information between the MPC controllers. Hence, the proposed DMPC strategy copes with this feature, resulting in a hierarchical-like scheme. Results obtained when used selected simulation scenarios show the effectiveness of the control strategy in terms of system modularity, reduced computational burden and, at the same time, the admissible loss of performance in contrast to a centralised MPC (CMPC) strategy., This work has been supported by Spanish research project WATMAN (CICYT DPI2009-13744) of the Science and Technology Ministry, the Juan de la Cierva Research Programme (ref. JCI-2008-2438), the DGR of Generalitat de Catalunya (SAC group Ref. 2009/SGR/1491) and the EU project WIDE (FP7-IST-224168), Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Contribución de congreso (ConferenceOutput)

Tuning of predictive controllers for drinking water networked systems

  • Toro, Rodrigo
  • Ocampo-Martínez, Carlos
  • Logist, Filip
  • Impe, Jan van
  • Puig, Vicenç
Trabajo presentado al 18th IFAC World Congress celebrado en Milán del 28 de agosto al 2 de septiembre de 2011., In this paper, two tuning strategies for a multi-objective predictive controller applied to a drinking water network (DWN) are proposed. A control-oriented DWN model is briefly reviewed, together with its management objectives. A comparison of methods to explore the Pareto front of the multi-objective optimisation (MOO) problem behind the predictive controller is presented with an eective normalisation method for the model predictive control (MPC) objectives. The proposed tuning strategies, applied to a real-life case study, are compared. Finally, simulation results show that the proposed MPC tuning strategies outperform the baseline results., Work of R. Toro, C. Ocampo-Martinez and V. Puig has been supported by the Spanish research project WATMAN (CICYT DPI2009-13744) of the Science and Technology Ministry, the DGR of Generalitat de Catalunya (SAC group Ref. 2009/SGR/1491) and the EU project WIDE (FP7-IST-224168). The authors specially thank the support received from AGBAR in the case study of this paper. Work of F. Logist and J.F.M. Van Impe is supported in part by Projects OT/09/025/TBA, OT/10/035, OPTEC (Center-of-Excellence Optimization in Engineering) PFV/10/002 and SCORES4CHEM KP/09/005 of the Katholieke Universiteit Leuven, and by the Belgian Program on Interuniversity Poles of Attraction, initiated by the Belgian Federal Science Policy Oce. J.F.M. Van Impe holds the chair Safety Engineering sponsored by the Belgian chemistry and life sciences federation essenscia., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Decentralised MPC based on a graph partitioning approach applied to the Barcelona drinking water network

  • Ocampo-Martínez, Carlos|||0000-0001-9251-6044
  • Puig Cayuela, Vicenç|||0000-0002-6364-6429
  • Bovo, Samuele
In this paper, the application of decentralised model predictive control (DMPC) over a partitioned model of the Barcelona drinking water network (DWN) is proposed. The system decomposition is based on graph partitioning theory, which considers the graph representation of the DWN topology. Once obtained, this system decomposition yields to some bidirectional flows of information between the MPC controllers. Hence, the proposed DMPC strategy copes with this feature, resulting in a hierarchical-like scheme. Results obtained when used selected simulation scenarios show the effectiveness of the control strategy in terms of system modularity, reduced computational burden and, at the same time, the admissible loss of performance in contrast to a centralised MPC (CMPC) strategy., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Decentralized model predictive control of drinking water networks using an automatic subsystem decomposition approach

  • Barcelli, Davide
  • Ocampo-Martínez, Carlos|||0000-0001-9251-6044
  • Puig Cayuela, Vicenç|||0000-0002-6364-6429
  • Bemporad, Alberto
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Negotiation and learning in distributed MPC of large scale systems

  • Javalera Rincón, Valeria
  • Morcego Seix, Bernardo|||0000-0002-6944-7519
  • Puig Cayuela, Vicenç|||0000-0002-6364-6429
A key issue in distributed MPC control of Large Scale Systems (LSS) is how shared variables among the different MPC controller in charge of controlling each system partition (subsystems) are handled. When these connections represent control variables, the distributed control has to be consistent for both subsystems and the optimal value of these variables will have to accomplish a common goal. In order to achieve this, the present work combines ideas from Distributed Artificial Intelligence (DAI), Reinforcement Learning (RL) and Model Predictive Control (MPC) in order to provide an approach based on negotiation, cooperation and learning techniques. Results of the application of this approach to a small drinking water network show that the resulting trajectories of the levels in tanks (control variables) can be acceptable compared to the centralized solution. The application to a real network (the Barcelona case) is currently under development., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Tuning of predictive controllers for drinking water networked systems

  • Toro Olmedo, Rodrigo
  • Ocampo-Martínez, Carlos|||0000-0001-9251-6044
  • Logist, Filip
  • Van Impe, J.
  • Puig Cayuela, Vicenç|||0000-0002-6364-6429
In this paper, two tuning strategies for a multi-objective predictive controller applied to a drinking water network (DWN) are proposed. A control-oriented DWN model is brie y reviewed, together with its management objectives. A comparison of methods to explore the Pareto front of the multi-objective optimisation (MOO) problem behind the predictive controller is presented with an e ective normalisation method for the model predictive control (MPC) objectives. The proposed tuning strategies, applied to a real-life case study, are compared. Finally, simulation results show that the proposed MPC tuning strategies outperform the baseline results., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

A multi-agent MPC architecture for distributed large scale systems

  • Javalera Rincón, Valeria
  • Morcego Seix, Bernardo|||0000-0002-6944-7519
  • Puig Cayuela, Vicenç|||0000-0002-6364-6429
In the present work, techniques of Model Predictive Control (MPC), Multi Agent Systems (MAS) and Reinforcement Learning (RL) are combined to develop a distributed control architecture for Large Scale Systems (LSS). This architecture is multi-agent based. The system to be controlled is divided in several partitions and there is an MPC Agent in charge of each partition. MPC Agents interact over a platform that allows them to be located physically apart. One of the main new concepts of this architecture is the Negotiator Agent. Negotiator Agents interact with MPC Agents which share control variables. These shared variables represent physical connections between partitions that should be preserved in order to respect the system structure. The case of study, in which the proposed architecture is being applied and tested, is a small drinking water network. The application to a real network (the Barcelona case) is currently under development.
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Distributed MPC for large scale systems using agent-based reinforcement learning

  • Javalera Rincón, Valeria
  • Morcego Seix, Bernardo|||0000-0002-6944-7519
  • Puig Cayuela, Vicenç|||0000-0002-6364-6429
In the present work, distributed control and artificial intelligence are combined in a control architecture for Large Scale Systems (LSS). The aim of this architecture is to provide a general structure and methodology to perform optimal control in networked distributed environments where multiple dependencies between sub-systems are found. Often these dependencies or connections represent control variables so the distributed control has to be consistent for both subsystems and the optimal value of these variables has to accomplish a common goal. The aim of the research described in this paper is to exploit the attractive features of MPC (meaningful objective functions and constraints) in a distributed implementation combining learning techniques to perform the negotiation of these variables in a cooperative Multi Agent environment and over a Multi Agent platform to provide speed, scalability, and computational effort reduction. This approach is based on negotiation, cooperation and learning. Results of the application of this architecture to a small drinking water network show that the resulting trajectories of the levels in tanks (control variables) can be acceptable compared to the centralized solution. The application to a real network (the Barcelona case) is currently under development.
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Model predictive control of drinking water networks: a hierarchical and decentralized approach

  • Ocampo-Martínez, Carlos|||0000-0001-9251-6044
  • Fambrini, Valentina
  • Barcelli, Davide
  • Puig Cayuela, Vicenç|||0000-0002-6364-6429
In this paper, a decentralized model predictive control (DMPC) strategy for drinking water networks (DWN)is proposed. The DWN is partitioned in a set of subnetworks using a partitioning algorithm that makes use of the topology of the network, the information about the actuator usage and heuristics. A suboptimal DMPC strategy was derived that allows the hierarchical solution of the set of MPC controllers used to control each partition. A comparative study between the centralized MPC (CMPC) and DMPC approaches is developed on the case study, which consists in an aggregate version of the Barcelona DWN. Results have shown the effectiveness of the proposed DMPC approach in terms of the computation time while an admissible level of suboptimality is obtained in all the considered scenarios., Peer Reviewed
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Decentralised MPC based on a graph partitioning approach applied to the Barcelona drinking water network

  • Ocampo-Martínez, Carlos
  • Puig Cayuela, Vicenç
  • Bovo, Samuele
In this paper, the application of decentralised model predictive control (DMPC) over a partitioned model of the Barcelona drinking water network (DWN) is proposed. The system decomposition is based on graph partitioning theory, which considers the graph representation of the DWN topology. Once obtained, this system decomposition yields to some bidirectional flows of information between the MPC controllers. Hence, the proposed DMPC strategy copes with this feature, resulting in a hierarchical-like scheme. Results obtained when used selected simulation scenarios show the effectiveness of the control strategy in terms of system modularity, reduced computational burden and, at the same time, the admissible loss of performance in contrast to a centralised MPC (CMPC) strategy., Peer Reviewed, Postprint (updated version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Negotiation and learning in distributed MPC of large scale systems

  • Javalera Rincón, Valeria
  • Morcego Seix, Bernardo
  • Puig Cayuela, Vicenç
A key issue in distributed MPC control of Large Scale Systems (LSS) is how shared variables among the different MPC controller in charge of controlling each system partition (subsystems) are handled. When these connections represent control variables, the distributed control has to be consistent for both subsystems and the optimal value of these variables will have to accomplish a common goal. In order to achieve this, the present work combines ideas from Distributed Artificial Intelligence (DAI), Reinforcement Learning (RL) and Model Predictive Control (MPC) in order to provide an approach based on negotiation, cooperation and learning techniques. Results of the application of this approach to a small drinking water network show that the resulting trajectories of the levels in tanks (control variables) can be acceptable compared to the centralized solution. The application to a real network (the Barcelona case) is currently under development., Peer Reviewed, Postprint (author’s final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Decentralized model predictive control of drinking water networks using an automatic subsystem decomposition approach

Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Model predictive control of drinking water networks: a hierarchical and decentralized approach

  • Ocampo-Martínez, Carlos
  • Fambrini, Valentina
  • Barcelli, Davide
  • Puig Cayuela, Vicenç
In this paper, a decentralized model predictive control (DMPC) strategy for drinking water networks (DWN)is proposed. The DWN is partitioned in a set of subnetworks using a partitioning algorithm that makes use of the topology of the network, the information about the actuator usage and heuristics. A suboptimal DMPC strategy was derived that allows the hierarchical solution of the set of MPC controllers used to control each partition. A comparative study between the centralized MPC (CMPC) and DMPC approaches is developed on the case study, which consists in an aggregate version of the Barcelona DWN. Results have shown the effectiveness of the proposed DMPC approach in terms of the computation time while an admissible level of suboptimality is obtained in all the considered scenarios., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Distributed MPC for large scale systems using agent-based reinforcement learning

  • Javalera Rincón, Valeria
  • Morcego Seix, Bernardo
  • Puig Cayuela, Vicenç
In the present work, distributed control and artificial intelligence are combined in a control architecture for Large Scale Systems (LSS). The aim of this architecture is to provide a general structure and methodology to perform optimal control in networked distributed environments where multiple dependencies between sub-systems are found. Often these dependencies or connections represent control variables so the distributed control has to be consistent for both subsystems and the optimal value of these variables has to accomplish a common goal. The aim of the research described in this paper is to exploit the attractive features of MPC (meaningful objective functions and constraints) in a distributed implementation combining learning techniques to perform the negotiation of these variables in a cooperative Multi Agent environment and over a Multi Agent platform to provide speed, scalability, and computational effort reduction. This approach is based on negotiation, cooperation and learning. Results of the application of this architecture to a small drinking water network show that the resulting trajectories of the levels in tanks (control variables) can be acceptable compared to the centralized solution. The application to a real network (the Barcelona case) is currently under development., Preprint
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Tuning of predictive controllers for drinking water networked systems

  • Toro Olmedo, Rodrigo
  • Ocampo-Martínez, Carlos
  • Logist, Filip
  • Van Impe, J.
  • Puig Cayuela, Vicenç
In this paper, two tuning strategies for a multi-objective predictive controller applied to a drinking water network (DWN) are proposed. A control-oriented DWN model is brie y reviewed, together with its management objectives. A comparison of methods to explore the Pareto front of the multi-objective optimisation (MOO) problem behind the predictive controller is presented with an e ective normalisation method for the model predictive control (MPC) objectives. The proposed tuning strategies, applied to a real-life case study, are compared. Finally, simulation results show that the proposed MPC tuning strategies outperform the baseline results., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

A multi-agent MPC architecture for distributed large scale systems

  • Javalera Rincón, Valeria
  • Morcego Seix, Bernardo
  • Puig Cayuela, Vicenç
In the present work, techniques of Model Predictive Control (MPC), Multi Agent Systems (MAS) and Reinforcement Learning (RL) are combined to develop a distributed control architecture for Large Scale Systems (LSS). This architecture is multi-agent based. The system to be controlled is divided in several partitions and there is an MPC Agent in charge of each partition. MPC Agents interact over a platform that allows them to be located physically apart. One of the main new concepts of this architecture is the Negotiator Agent. Negotiator Agents interact with MPC Agents which share control variables. These shared variables represent physical connections between partitions that should be preserved in order to respect the system structure. The case of study, in which the proposed architecture is being applied and tested, is a small drinking water network. The application to a real network (the Barcelona case) is currently under development., Postprint (published version)