Optimal solutions of selected cellular neural network applications by the hardware annealing method

Author

Sheu, Bing J. ; Bang, Sa H. ; Fang, Wai-Chi

Author_Institution

Dept. of Electr. Eng., Univ. of Southern California, Los Angeles, CA, USA

fYear

1994

fDate

18-21 Dec 1994

Firstpage

279

Lastpage

284

Abstract

An engineering annealing method for optimal solutions of cellular neural networks is presented. Cellular neural networks have great potential in solving many important scientific problems in signal processing and optimization by the use of predetermined templates. Hardware annealing (SaHyun Bang, 1994), which is a paralleled version of effective mean field annealing in analog networks, is a highly efficient method of finding optimal solutions for cellular neural networks. It does not require any stochastic procedure and henceforth can be very fast. The generalized energy function of the network is first increased by reducing the voltage gain of each neuron. Then, the hardware annealing searches for the globally minimum energy state by continuously increasing the gain of neurons. The process of global optimization by the proposed hardware annealing method can be described by eigenvalues in the time varying dynamic system

Keywords

bioelectric phenomena; cellular neural nets; neural chips; optimisation; cellular neural network applications; effective mean field annealing; eigenvalues; engineering annealing method; generalized energy function; globally minimum energy state; hardware annealing method; optimal solutions; predetermined templates; time varying dynamic system; voltage gain reduction; Annealing; Cellular neural networks; Eigenvalues and eigenfunctions; Energy states; Neural network hardware; Neurons; Optimization methods; Signal processing; Stochastic processes; Voltage;

fLanguage

English

Publisher

ieee

Conference_Titel

Cellular Neural Networks and their Applications, 1994. CNNA-94., Proceedings of the Third IEEE International Workshop on

Conference_Location

Rome

Print_ISBN

0-7803-2070-0

Type

conf

DOI

10.1109/CNNA.1994.381666

Filename

381666