cmu: CMU data from Blocker & Airoldi (2011)

In awblocker/networkTomography: Tools for Network Tomography

CMU data from Blocker & Airoldi (2011)

Description

Data from the 12 node CMU network used in Blocker & Airoldi (2011). The OD flows are actual, observed traffic from a CMU network. The topology does not, however, correspond to the original network due to security considerations.

Usage

data(cmu)

Objects

The list cmu, which contains several objects:

• A, the routing matrix for this network (truncated for full row rank)

• X, a matrix of origin-destination flows formatted for analysis

• Y, a matrix of link loads formatted for analysis

• A.full, the routing matrix for this network without truncation for full row rank)

• Y.full, a matrix of link loads corresponding to codeA.full

In this data, we have A %*% t(X) == t(Y) and A.full %*% t(X) == t(Y.full)

Variables

The list cmu contains the following:

• The routing matrix A. The columns of this matrix correspond to individual OD flows (the columns of X), and its rows correspond to individual link loads (the columns of Y).

• The OD matrix X. Columns correspond to individual OD flows, and the rows correspond to observations.

• The link load matrix Y. Columns of the Y matrix correspond to individual link loads, and the rows correspond to observations.

• The routing matrix A.full. This is the complete routing matrix before reduction for full row-rank.

• The link load matrix Y.full, corresponding to A.full.

References

A.W. Blocker and E.M. Airoldi. Deconvolution of mixing time series on a graph. Proceedings of the Twenty-Seventh Conference Annual Conference on Uncertainty in Artificial Intelligence (UAI-11) 51-60, 2011.

awblocker/networkTomography documentation built on May 14, 2022, 10:05 p.m.