TY - GEN
T1 - Exploring Koopman Operator Based Surrogate Models—Accelerating the Analysis of Critical Pedestrian Densities
AU - Lehmberg, Daniel
AU - Dietrich, Felix
AU - Kevrekidis, Ioannis G.
AU - Bungartz, Hans Joachim
AU - Köster, Gerta
N1 - Funding Information:
Acknowledgments This work is supported by the German Research Foundation (DFG) grant no. KO 5257/3-1. The work of I.G.K. was partially supported by the DARPA PAI program. D.L. thanks the research office (FORWIN) of Munich University of Applied Sciences and the Faculty Graduate Center CeDoSIA of TUM Graduate School at Technical University of Munich for their support.
Publisher Copyright:
© 2020, Springer Nature Switzerland AG.
PY - 2020
Y1 - 2020
N2 - We apply the Koopman operator framework to pedestrian dynamics. In an example scenario, we generate crowd density time series data with a microscopic pedestrian simulator. We then approximate the Koopman operator in matrix form through Extended Dynamic Mode Decomposition, using Geometric Harmonics on the data as a dictionary. The Koopman matrix is integrated into a surrogate model, which allows to approximate crowd density time series data to be generated, independently from the original microscopic simulator. The evaluation of the constructed surrogate model is orders of magnitude faster, and enables us to use methods that require many model evaluations.
AB - We apply the Koopman operator framework to pedestrian dynamics. In an example scenario, we generate crowd density time series data with a microscopic pedestrian simulator. We then approximate the Koopman operator in matrix form through Extended Dynamic Mode Decomposition, using Geometric Harmonics on the data as a dictionary. The Koopman matrix is integrated into a surrogate model, which allows to approximate crowd density time series data to be generated, independently from the original microscopic simulator. The evaluation of the constructed surrogate model is orders of magnitude faster, and enables us to use methods that require many model evaluations.
UR - http://www.scopus.com/inward/record.url?scp=85097293507&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85097293507&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-55973-1_19
DO - 10.1007/978-3-030-55973-1_19
M3 - Conference contribution
AN - SCOPUS:85097293507
SN - 9783030559724
T3 - Springer Proceedings in Physics
SP - 149
EP - 157
BT - Traffic and Granular Flow 2019
A2 - Zuriguel, Iker
A2 - Garcimartín, Angel
A2 - Hidalgo, Raúl Cruz
PB - Springer Science and Business Media Deutschland GmbH
T2 - 13th Conference on Traffic and Granular Flow, TGF 2019
Y2 - 2 July 2019 through 5 July 2019
ER -