Adjoint Sensitivity Analysis for Differential-Algebraic Equations

SHENGTAI LI
Computer Science Department
University of Southern California Santa Barbara
Santa Barbara, CA 93106
shengtai@cs.ucsb.edu

Abstract:  We have developed a robust, efficient numerical method for sensitivity computation of large-scale differential-algebraic equation (DAE) systems. The numerical method calculates the sensitivities of an initial value problem (IVP) by solving the adjoint DAE system backward. We also address the well-posedness of sensitivity problems for DAEs and the numerical stability for the adjoint system of the BDF methods. Complexity analysis and numerical results show that the adjoint sensitivity method is advantageous over the forward sensitivity method for applications with a large number of sensitivity parameters and few derived functions.