http://www.rssboard.org/rss-specification 720 https://escholarship.org/uc/ucistat_papers/rss Recent eScholarship items from Faculty Publications Sat, 27 Jun 2026 00:58:45 +0000 https://escholarship.org/uc/item/7q18k8v9 <p>In this paper, we propose an approach to estimating traffic matrices that incorporates lightweight Origin- Destination (OD) flow measurements coupled with a computationally lightweight algorithm for producing the OD estimates. There are two key ingredients in our method, called PamTram, for PArtial Measurement of TRAffic Matrices. The first is to actively select a small number of informative OD flows to measure in each estimation time interval. To avoid the heavy computation of an optimal selection, we use a heuristic based on intuition from game theory. Randomized selection rules are developed based on the goals of reducing errors and adapting to traffic changes. We provide an algorithm for selecting a good flow to measure that is fast because it avoids the computations, such as integrating over past intervals, that are needed for optimal selection. The second key aspect of our method is an explanation and proof that an Iterative Proportional Fitting (IPF) algorithm can be... https://escholarship.org/uc/item/7q18k8v9 Fri, 3 Jun 2005 00:00:00 +0000 Liang, Gang Taft, Nina Yu, Bin