In this article we address the model order reduction problem for resistor networks by using methods from graph theory. We formulate this problem through graph theory concepts, propose algorithms for solving it, and present the computational results we have obtained for real-world resistor networks. The results demonstrate that graph-theoretical methods produce networks that contain fewer edges and are sparser than networks produced by state-of-the-art methods.
Published April 2011 , 19 pages
This cahier was revised in October 2011