This paper presents a stochastic algorithm which is commonly used in signal and image processing. We show that simulated annealing is a powerful minimization method to avoid local minima and find the global minimum of a function. We control three critical parameters to obtain the algorithm convergence: an initial temperature, a mathematical equilibrium criterion and a cooling factor. We illustrate these results with a synthesis of a pass-band filter. We emphasize the dependence of the algorithm performance on the cooling factor.

Date Published: 4 November 1994
PDF: 8 pages
Proc. SPIE 2253, Optical Interference Coatings, (4 November 1994); doi: 10.1117/12.192173

Published in SPIE Proceedings Vol. 2253:
Optical Interference Coatings
Florin Abeles, Editor(s)