Storage and transmission of signals tend to be resource demanding when the data are processed in their original size. Compressing the signal in such a way that close reproductions can be found, is therefore a highly respected problem in signal processing. In this talk, we demonstrate how optimization methods can contribute to this, and we pay particular attention to the compression of one-dimensional signals such as electro cardiograms (ECG).
Applications to ECG signals require response from the compression algorithm in real time. We show how the compression problem can be modeled as a constrained shortest path problem, and we suggest a simple polynomial-time algorithm designed in order to respect the real-time requirement. We study the computational properties of the algorithm, and we point out some future perspectives of the method.