A randomized proper orthogonal decomposition technique

In this paper, we consider the problem of model reduction of large scale systems, such as those obtained through the discretization of PDEs. We propose a randomized proper orthogonal decomposition (RPOD) technique to obtain the reduced order models by randomly choosing a subset of the inputs/outputs of the system to construct a suitable small sized Hankel matrix from the full Hankel matrix. It is shown that the RPOD technique is computationally orders of magnitude cheaper when compared to techniques such as the Eigensystem Realization Algorithm (ERA)/Balanced proper orthogonal decomposition (BPOD) while obtaining the same information in terms of the number and accuracy of the dominant modes. The method is tested on a linearized channel flow problem.