An Interactive Genetic Algorithm for Controller Parameter Optimization
Genetic algorithms are stochastic search algorithms inspired by biological phenomena of genetic recombination and natural selection. They simulate the evolution of string individuals encoding candidate solutions to a given problem. Genetic algorithms proved robust and efficient in finding near-optimal solutions in complex problem spaces. They are usually exploited as an optimization method, suitable for both continuous and discrete optimization tasks.
In this paper, genetic algorithms are investigated as an engineering optimization tool. The work focuses on tuning parameters in control system design. This domain has already been approached with genetic algorithms, but most of the experiments have been done using computer simulations of the devices to be controlled. We present the Interactive Genetic Algorithm (IGA) for controller parameter optimization. IGA carries out the evolution of parameter values in a traditional manner, but differs from a conventional genetic algorithm in that it allows interaction with real-world environment. The algorithm suggests the trials to be performed to explore the parameter space, and accepts results of the trials from the environment. The paper describes IGA and its application in tuning the parameters of a PID regulator operating on a laboratory device.
KeywordsGenetic Algorithm Candidate Solution Settling Time Binary String Control System Design
Unable to display preview. Download preview PDF.
- Holland J. H., ‘Adaptation in Natural and Artificial Systems’, University of Michigan Press, Ann Arbor, 1975.Google Scholar
- Goldberg D. E., ‘Genetic Algorithms in Search, Optimization and Machine Learning’, Addison-Wesley, Reading, 1989.Google Scholar
- Rawlins G. J. E., ‘Foundations of Genetic Algorithms’, Morgan Kaufmann, San Mateo, 1991.Google Scholar
- Davis L. (ed.), ‘Handbook of Genetic Algorithms’, Van Nostrand Reinhold, New York, 1991.Google Scholar
- Goldberg D. E., ‘Dynamic system control using rule learning and genetic algorithms’, Proc. 9th Int. Joint Conf. on Artificial Intelligence, Los Angeles, 1985, pp. 588–592.Google Scholar
- Odetayo M. O. and McGregor D. R., ‘Genetic algorithm for inducing control rules for a dynamic system’, Proc. 3rd Int. Conf. on Genetic Algorithms, Morgan Kaufmann, San Mateo, 1989, pp. 177–182.Google Scholar
- Varšek A., Urbančič T. and Filipič B., ‘Genetic algorithms in controller design and tuning’, To appear in IEEE Trans. Syst., Man and Cybernetics, 1993.Google Scholar
- Hunt K. J., ‘Optimal control system synthesis with genetic algorithms’, Männer R. and Manderick B. (eds.), Parallel Problem Solving from Nature 2, Elsevier, Amsterdam, 1992, pp. 381-389.Google Scholar
- Hunt K. J.,—, ‘Preprints IFAC/IFIP/IMACS Int. Symp. on Artificial Intelligence in Real-Time Control’, Delft, The Netherlands, 1992, Session 31: Genetic algorithms and learning, pp. 579-610.Google Scholar
- Ziegler J. G. and Nichols N. B., ‘Optimum settings for automatic controllers’, Trans. A.S.M.E., Vol. 64, 1942, pp. 759–768.Google Scholar
- Zupančič B., ‘SIMCOS Language for simulation of continuous and discrete dynamic systems’, Faculty of Electrical Engineering and Computer Science, University of Ljubljana, 1991.Google Scholar
- Urbančič T., Juričić D., Filipič B. and Bratko I., ‘Automated synthesis of control for nonlinear dynamic systems’, Preprints IFAC/IFIP/IMACS Int. Symp. on Artificial Intelligence in Real-Time Control, Delft, The Netherlands, pp. 605–610.Google Scholar
- Pavlinič A., ‘Modelling and computer control of an unstable laboratory process’, B.Sc. Thesis, Faculty of Electrical Engineering and Computer Science, University of Ljubljana, 1991 (in Slovenian).Google Scholar