|
CMU-CS-01-37
Computer Science Department
School of Computer Science, Carnegie Mellon University
CMU-CS-01-137
Stochastic Search for Signal Processing Algorithm Optimization
Bryan Singer, Manuela Veloso
May 2001
CMU-CS-01-137.ps
CMU-CS-01-137.pdf
Keywords: Evolutionary algorithms, signal processing, performance
optimization, genertic algorithms
Many difficult problems can be viewed as search problems. However,
given a new task with an embedded search problem, it is challenging to
state and find a truly effective search approach. In this paper, we
address the complex task of signal processing optimization. We first
introduce and discuss the complexities of this domain. In general, a
single signal processing algorithm can be represented by a very large
number of different but mathematically equivalent formulas. When these
formulas are implemented in actual code, unfortunately their running
times differ significantly. Signal processing algorithm optimization
aims at finding the fastest formula. We present a new approach that
successfully solves this problem, using an evolutionary stochastic
search algorithm, STEER, to search through the very large space of
formulas. We empirically compare STEER against other search methods,
showing that it notably can find faster formulas while still only timing
a very small portion of the search space.
13 pages
|