Title :
Optimizing neural network for searching of function extrema in the Remez algorithm
Author_Institution :
Inst. of Telecommun., Wroclaw Univ. of Technol., Wroclaw
Abstract :
In this paper, a general algorithm for finding of function extrema in the Remez algorithm is presented. The proposed algorithm utilizes neural network, gradient method and least square method which allows one to achieve precise solution of the problem. Moreover since neural network allows a parallel solving for several points, it means the proposed algorithm is time efficient. The gradient method and the least square method make the algorithm easy to implement.
Keywords :
FIR filters; gradient methods; least squares approximations; neural nets; Remez algorithm; function extrema; gradient method; least square method; neural network; Acoustics; Algorithm design and analysis; Digital filters; Finite impulse response filter; Gradient methods; Iterative algorithms; Iterative methods; Least squares methods; Neural networks; Signal processing algorithms; Remez algorithm; digital filter design; neural networks; optimization;
Conference_Titel :
Signals and Electronic Systems, 2008. ICSES '08. International Conference on
Conference_Location :
Krakow
Print_ISBN :
978-83-88309-47-2
Electronic_ISBN :
978-83-88309-52-6
DOI :
10.1109/ICSES.2008.4673419
