Julian Alarcon - Academia.edu (original) (raw)
Papers by Julian Alarcon
Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference, 2009
... Julián Bonilla, Member, IEEE, Moritz Diehl, Filip Logist, Bart De Moor Fellow, IEEE, and Jan ... more ... Julián Bonilla, Member, IEEE, Moritz Diehl, Filip Logist, Bart De Moor Fellow, IEEE, and Jan Van Impe ... julian.bonilla@cit.kuleuven.be M. Diehl is an associate professor at the Department of Electrical Engineering ESAT/SCD, Katholieke Universiteit Leuven, B-3001 Leuven ...
2008 47th IEEE Conference on Decision and Control, 2008
This paper introduces a convex formulation approach for the initialization of parameter estimatio... more This paper introduces a convex formulation approach for the initialization of parameter estimation problems (PEP). The proposed method exploits the parameter-affine feature exhibited by some dynamic systems. The method attempts to solve a related convex problem and uses its result as the initial guess for the solution of the original nonconvex PEP. The proposed approach is illustrated through two nonconvex parameter estimation study cases; the harmonic oscillator and the Lorenz attractor. I. INTRODUCTION Obtaining accurate models of dynamic systems has an enormous impact on science and engineering. Models used to predict and control process dynamics are basically characterized by their structure and the parameter values in this structure. Parameter estimation problems refer to the calculation of a set of values in a predefined model structure, linear or nonlinear, such that the outputs of the model for this given set of parameters fit some measurement data. Additionally, constraints on parameter values and model states are usually required e.g. positive concentrations and reaction rates, upper and lower limits in model outputs. Consequently, PEP are recast as optimization problems leading to convex or nonconvex formulations according to the nature of the cost, model and constraints. Nonconvex PEP are difficult to solve since they might exhibit local solutions and the true parameter or global solution might be hard to find. Fast and efficient techniques based on Newton type methods have been proposed [1]. These methods are derivative-based, and they can easily lock on to a local solution if the problem is not initialized appropriately. The most reliable approches for PEP are based on the constrained Gauss-Newton method, with simultaneous optimization [2]. Nevertheless, all these approaches require a starting point to initialize the algorithms. Other optimization techniques have been also introduced for these kind of problems, such as non-deterministic global optimization methods based on random search, genetic algorithms and simulated annealing and deterministic approaches
Optimal Control Applications and Methods, 2010
This paper presents a convexity-based homotopy solution procedure to non-convex optimal control p... more This paper presents a convexity-based homotopy solution procedure to non-convex optimal control problems (OCPs) arising in model predictive control. The approach deals with a special class of OCP formulations where the dynamic system involved is control-affine and the objective is a penalty on deviations from a state reference trajectory. The non-convex OCP is modified by introducing a penalized pseudo state and a homotopy parameter which gradually transforms the original problem into a convex one. The method solves first this convex formulation and uses the result to initialize the solution of the next problem on the zero path, recovering the original OCP. The proposed methodology is evaluated for the benchmark control problem of an isothermal chemical reactor with Van de Vusse reactions and input multiplicity. For the simple case with control horizon one, the method is able to find the global solution due to the convex initialization, while local optimization techniques only lead to a local minimum.
Computers & Chemical Engineering, 2010
This document contains the post-print pdf-version of the refereed paper: "An automatic initializa... more This document contains the post-print pdf-version of the refereed paper: "An automatic initialization procedure in parameter estimation problems with parameter-affine dynamic models"
Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference, 2009
... Julián Bonilla, Member, IEEE, Moritz Diehl, Filip Logist, Bart De Moor Fellow, IEEE, and Jan ... more ... Julián Bonilla, Member, IEEE, Moritz Diehl, Filip Logist, Bart De Moor Fellow, IEEE, and Jan Van Impe ... julian.bonilla@cit.kuleuven.be M. Diehl is an associate professor at the Department of Electrical Engineering ESAT/SCD, Katholieke Universiteit Leuven, B-3001 Leuven ...
2008 47th IEEE Conference on Decision and Control, 2008
This paper introduces a convex formulation approach for the initialization of parameter estimatio... more This paper introduces a convex formulation approach for the initialization of parameter estimation problems (PEP). The proposed method exploits the parameter-affine feature exhibited by some dynamic systems. The method attempts to solve a related convex problem and uses its result as the initial guess for the solution of the original nonconvex PEP. The proposed approach is illustrated through two nonconvex parameter estimation study cases; the harmonic oscillator and the Lorenz attractor. I. INTRODUCTION Obtaining accurate models of dynamic systems has an enormous impact on science and engineering. Models used to predict and control process dynamics are basically characterized by their structure and the parameter values in this structure. Parameter estimation problems refer to the calculation of a set of values in a predefined model structure, linear or nonlinear, such that the outputs of the model for this given set of parameters fit some measurement data. Additionally, constraints on parameter values and model states are usually required e.g. positive concentrations and reaction rates, upper and lower limits in model outputs. Consequently, PEP are recast as optimization problems leading to convex or nonconvex formulations according to the nature of the cost, model and constraints. Nonconvex PEP are difficult to solve since they might exhibit local solutions and the true parameter or global solution might be hard to find. Fast and efficient techniques based on Newton type methods have been proposed [1]. These methods are derivative-based, and they can easily lock on to a local solution if the problem is not initialized appropriately. The most reliable approches for PEP are based on the constrained Gauss-Newton method, with simultaneous optimization [2]. Nevertheless, all these approaches require a starting point to initialize the algorithms. Other optimization techniques have been also introduced for these kind of problems, such as non-deterministic global optimization methods based on random search, genetic algorithms and simulated annealing and deterministic approaches
Optimal Control Applications and Methods, 2010
This paper presents a convexity-based homotopy solution procedure to non-convex optimal control p... more This paper presents a convexity-based homotopy solution procedure to non-convex optimal control problems (OCPs) arising in model predictive control. The approach deals with a special class of OCP formulations where the dynamic system involved is control-affine and the objective is a penalty on deviations from a state reference trajectory. The non-convex OCP is modified by introducing a penalized pseudo state and a homotopy parameter which gradually transforms the original problem into a convex one. The method solves first this convex formulation and uses the result to initialize the solution of the next problem on the zero path, recovering the original OCP. The proposed methodology is evaluated for the benchmark control problem of an isothermal chemical reactor with Van de Vusse reactions and input multiplicity. For the simple case with control horizon one, the method is able to find the global solution due to the convex initialization, while local optimization techniques only lead to a local minimum.
Computers & Chemical Engineering, 2010
This document contains the post-print pdf-version of the refereed paper: "An automatic initializa... more This document contains the post-print pdf-version of the refereed paper: "An automatic initialization procedure in parameter estimation problems with parameter-affine dynamic models"