DocumentCode :
1100640
Title :
The design of optimal DFT algorithms using dynamic programming
Author :
Johnson, Howard W. ; Burrus, C. Sidney
Author_Institution :
ROLM Corporation, Santa Clara, CA
Volume :
31
Issue :
2
fYear :
1983
fDate :
4/1/1983 12:00:00 AM
Firstpage :
378
Lastpage :
387
Abstract :
A broad class of efficient discrete Fourier transform algorithms is developed by partitioning short DFT algorithms into factors. The factored short DFT´s are combined into longer DFT´s using multi-dimensional index maps. By exploiting a property which allows some of the factors to commute, a large set of possible DFT algorithms is generated which contains both the prime factor algorithm (PFA) and the Winograd Fourier transform algorithm (WFTA) as special cases. The problem of finding an algorithm from this class which is optimal with respect to the specific add, multiply, and data transfer characteristics of a particular implementation is posed, and a highly effective dynamic programming algorithm is presented as a solution.
Keywords :
Algorithm design and analysis; Constraint optimization; Discrete Fourier transforms; Dynamic programming; Fourier transforms; Hardware; Heuristic algorithms; Indexing; Matrix decomposition; Partitioning algorithms;
fLanguage :
English
Journal_Title :
Acoustics, Speech and Signal Processing, IEEE Transactions on
Publisher :
ieee
ISSN :
0096-3518
Type :
jour
DOI :
10.1109/TASSP.1983.1164071
Filename :
1164071
Link To Document :
بازگشت