Abstract
Adaptive and learning control systems normally require methods which search automatically for the set of parameters which optimize a performance index. Since environmental or other conditions may change in a learning situation, the optimization must occur continuously and adapt to the existing conditions. If the performance index is unimodal and noisy, but non-stationary, a combination of gradient and random methods yields a search method which exhibits the advantages of both the random and stochastic approximation methods. The overall rate of convergence is improved and under certain conditions can be optimized. In addition, the random component of such a search provides a capability for multimodal search applications.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
J. Blum, “Multidimensional Stochastic Approximation Method”, Anals of Math. Stat., 25, pp. 400–407 (1954).
S. H. Brooks, “A Discussion of Random Methods for Seeking Maxima”, Operations Research 6, No.2, pp. 244–251 (1958).
A. Dvoretsky, “On Stochastic Approximation”, Proc. 3rd Berkeley Symp. on Math. Stat. and Prob., 1956.
L. S. Gurin and L. A. Rastrigin, “Convergence of the Random Search Method in the Presence of Noise”, Automation and Remote Control 26, No.9, pp. 1505–1511 (1965).
A. N. Kabalevskii, “Analysis of Search Under Noisy Conditions by Means of Markov Processes”, Automation and Remote Control 26, No.11, pp. 1938–1946 (1965).
D. C. Karnopp, “Random Search Techniques for Optimization Problems”, Automatica 1, pp. 111–121 (1963).
J. Kiefer, and J. Wolfowitz, “Stochastic Estimation of the Maximum of a Regression function”, Annals of Math. Stat. 23, pp. 462–466 (1952).
J. Matyas, “Random Optimization”, Automation and Remote Control 26, No. 2, pp. 244–251 (1965).
G. J. McMurtry, Adaptive, Learning, and Pattern Recognition Systems, Mendel and Fu, ed., Academic Press, 1970.
A. F. Pensa, “Gradient Biased Random Search”, Ph.D. Thesis, Department of E.E., Penn State University, 1969.
A. F. Pensa and G. J. McMurtry, “Gradient Biased Random Search”, Proc. IEEE 1970 Sys. Science and Cybernetics Conf., 1970.
L. A. Rastrigin, “The onvergence of the Random Search Method in the Extremal Control of a Many-Parameter System”, Automation and Remote Control 24, No. 11, pp. 1337–1342 (1963).
E. M. Vaysbord and D. B. Yudin, “Multiextremal Stochastic Approximation”, Engineering Cybernetics 6, No. 5, pp. 1–11 (1968).
M. D. Waltz and K. S. Fu, “A Heuristic Approach to Learning Control Systems”, IEEE Trans. Auto. Control AC-10, No. 4, pp. 390–398 (1965).
D. J. Wilde, Optimum Seeking Methods, Prentice-Hall, Englewood Cliffs, New Jersey, 1964.
G. N. Saridis, “On a Class of Performance-Adaptive Self-Organizing Control Systems,” this volume.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1971 Plenum Press, New York
About this chapter
Cite this chapter
McMurtry, G.J. (1971). Adaptive Optimization in Learning Control. In: Fu, K.S. (eds) Pattern Recognition and Machine Learning. Springer, Boston, MA. https://doi.org/10.1007/978-1-4615-7566-5_16
Download citation
DOI: https://doi.org/10.1007/978-1-4615-7566-5_16
Publisher Name: Springer, Boston, MA
Print ISBN: 978-1-4615-7568-9
Online ISBN: 978-1-4615-7566-5
eBook Packages: Springer Book Archive