Asynchronous systems for constraint satisfaction: Filtering and stability

Tarek A. Lahlou; Thomas A. Baran

ICASSP 2016accepted0 citations

Asynchronous systems for constraint satisfaction: Filtering and stability

Abstract

We discuss the formulation of, and stability conditions for, a set of generally asynchronous signal processing systems for solving a class of constraint satisfaction problems. Problems within this class are specifically those where a quadratic conservation principle is known to exist, as commonly occurs in stationarity conditions associated with a variety of convex and nonconvex optimization problems. With the intent of addressing a wide range of system architectures, the presented stability results are formulated for use with both a generally asynchronous update protocol, modeled as sample-and-hold subsystems triggered by independent Bernoulli processes, and also first-order filtering of the asynchronous updates. Numerical examples are provided by illustration and reference, indicating system behavior consistent with the presented stability results.

BibTeX

@inproceedings{icassp2016_asynchronoussyst,
  title = {Asynchronous systems for constraint satisfaction: Filtering and stability},
  author = {Tarek A. Lahlou and Thomas A. Baran},
  booktitle = {ICASSP 2016},
  year = {2016}
}