Multi-objective Optimization of PID Controller using Pareto-based Surrogate Modeling Algorithm for MIMO Evaporator System

Most control engineering problems are characterized by several objectives, which have to be satisfied simultaneously. Two widely used methods for finding the optimal solution to such problems are aggregating to a single criterion, and using Pareto-optimal solutions. This paper proposed a Pareto-based Surrogate Modeling Algorithm (PSMA) approach using a combination of Surrogate Modeling (SM) optimization and Pareto-optimal solution to find a fixed-gain, discrete-time Proportional Integral Derivative (PID) controller for a Multi Input Multi Output (MIMO) Forced Circulation Evaporator (FCE) process plant. Experimental results show that a multi-objective, PSMA search was able to give a good approximation to the optimum controller parameters in this case. The Non-dominated Sorting Genetic Algorithm II (NSGA-II) method was also used to optimize the controller parameters and as comparison with PSMA.

[1]  T. Simpson,et al.  Comparative studies of metamodelling techniques under multiple modelling criteria , 2001 .

[2]  Kalyanmoy Deb,et al.  A fast and elitist multiobjective genetic algorithm: NSGA-II , 2002, IEEE Trans. Evol. Comput..

[3]  Amrul Faruq,et al.  Optimization of depth control for Unmanned Underwater Vehicle using surrogate modeling technique , 2011, 2011 Fourth International Conference on Modeling, Simulation and Applied Optimization.

[4]  M. Mohamed Sultan,et al.  CONTROLLERS OPTIMIZATION FOR A FLUID MIXING SYSTEM USING METAMODELLING APPROACH , 2009 .

[5]  Mohd Ashraf Ahmad,et al.  Optimization of PID controllers for cartesian coordinates control of hovercraft system using metamodeling approach , 2008 .

[6]  Shahrum Shah bin Abdullah,et al.  Multi-objective optimization of an evaporator control system using surrogate modeling , 2011, 2011 IEEE International Conference on Control System, Computing and Engineering.

[7]  Isabel R. Santos,et al.  Simulation metamodels for modeling output distribution parameters , 2007, 2007 Winter Simulation Conference.

[8]  R. B. Newell,et al.  Applied Process Control: A Case Study , 1989 .

[9]  Jack P. C. Kleijnen,et al.  A methodology for fitting and validating metamodels in simulation , 2000, Eur. J. Oper. Res..

[10]  David S. Broomhead,et al.  Multivariable Functional Interpolation and Adaptive Networks , 1988, Complex Syst..

[11]  Shahrum Shah bin Abdullah,et al.  Multi-objective optimization of remotely operated vehicle control system using surrogate modeling , 2011, 2011 IEEE International Conference on Control System, Computing and Engineering.

[12]  Kunlun Xin,et al.  Using Radial Basis Function Neural Networks to Calibrate Water Quality Model , 2008 .

[13]  Lothar Thiele,et al.  Comparison of Multiobjective Evolutionary Algorithms: Empirical Results , 2000, Evolutionary Computation.

[14]  S. S. Abdullah,et al.  Metamodeling approach for PID controller optimization in an evaporator process , 2011, 2011 Fourth International Conference on Modeling, Simulation and Applied Optimization.

[15]  Shahrum Shah Abdullah,et al.  An active learning approach for radial basis function neural networks , 2006 .