Title :
A time-unrolling method to compute sensitivity of dynamic systems
Author :
Liu, Frank ; Feldmann, Peter
Author_Institution :
IBM Res. Austin, Austin, TX, USA
Abstract :
Sensitivities of the dynamic system responses with respect to the system parameters are highly valuable, with broad applications such as system tuning and uncertainty quantification. Compared to the direct methods, adjoint methods are much more efficient when the number of parameters is large. In this paper, we present a time-unrolling method to compute adjoint sensitivities. Instead of explicitly constructing the adjoint system, which quite often is nontrivial, our time-unrolling method implicitly retrace the response trajectory by utilizing the fitting polynomial of the integration methods. This paper provides theoretical foundation of the method as well as experimental demonstrations of its effectiveness.
Keywords :
nonlinear systems; polynomials; sensitivity; sensitivity analysis; tuning; adjoint sensitivities; dynamic system sensitivity computation; fitting polynomial; integration methods; response trajectory; system parameters; system tuning; time-unrolling method; uncertainty quantification; Clocks; Jacobian matrices; Nonlinear dynamical systems; Polynomials; Sensitivity; Trajectory;
Conference_Titel :
Design Automation Conference (DAC), 2014 51st ACM/EDAC/IEEE
Conference_Location :
San Francisco, CA
DOI :
10.1145/2593069.2593080