Research Papers

The Discrete Adjoint Gradient Computation for Optimization Problems in Multibody Dynamics

[+] Author and Article Information
Thomas Lauß

University of Applied Sciences Upper Austria,
Stelzhamerstrae 23,
Wels 4600, Austria;
Institute of Mechanics and Mechatronics,
Vienna University of Technology,
Getreidemarkt 9/E325,
Wien 1060, Austria
e-mail: thomas.lauss@fh-wels.at

Stefan Oberpeilsteiner

University of Applied Sciences Upper Austria,
Stelzhamerstrae 23,
Wels 4600, Austria;
Institute of Mechanics and Mechatronics,
Vienna University of Technology,
Getreidemarkt 9/E325,
Wien 1060, Austria
e-mail: stefan.oberpeilsteiner@fh-wels.at

Wolfgang Steiner

University of Applied Sciences Upper Austria,
Stelzhamerstrae 23,
Wels 4600, Austria
e-mail: wolfgang.steiner@fh-wels.at

Karin Nachbagauer

University of Applied Sciences Upper Austria,
Stelzhamerstrae 23,
Wels 4600, Austria
e-mail: karin.nachbagauer@fh-wels.at

1Corresponding author.

Contributed by the Design Engineering Division of ASME for publication in the JOURNAL OF COMPUTATIONAL AND NONLINEAR DYNAMICS. Manuscript received July 15, 2016; final manuscript received October 24, 2016; published online December 5, 2016. Assoc. Editor: Paramsothy Jayakumar.

J. Comput. Nonlinear Dynam 12(3), 031016 (Dec 05, 2016) (10 pages) Paper No: CND-16-1336; doi: 10.1115/1.4035197 History: Received July 15, 2016; Revised October 24, 2016

The adjoint method is a very efficient way to compute the gradient of a cost functional associated to a dynamical system depending on a set of input signals. However, the numerical solution of the adjoint differential equations raises several questions with respect to stability and accuracy. An alternative and maybe more natural approach is the discrete adjoint method (DAM), which constructs a finite difference scheme for the adjoint system directly from the numerical solution procedure, which is used for the solution of the equations of motion. The method delivers the exact gradient of the discretized cost functional subjected to the discretized equations of motion. For the application of the discrete adjoint method to the forward solver, several matrices are necessary. In this contribution, the matrices are derived for the simple Euler explicit method and for the classical implicit Hilber–Hughes–Taylor (HHT) solver.

Copyright © 2017 by ASME
Your Session has timed out. Please sign back in to continue.


Lauss, T. , Leitner, P. , Oberpeilsteiner, S. , and Steiner, W. , 2015, “ Energy Optimal Manipulation of an Industrial Robot,” ECCOMAS Thematic Conference on Multibody Dynamics, Barcelona, Catalonia, Spain, June 29–July 2, pp. 789–798. http://congress.cimne.com/multibody2015/admin/files/fileabstract/a84.pdf
Bryson, A. , and Ho, Y. , 1975, Applied Optimal Control, Hemisphere, Washington, DC.
Haug, E. , Wehage, R. , and Mani, N. , 1984, “ Design Sensitivity Analysis of Large-Scaled Constrained Dynamic Mechanical Systems,” J. Mech., Transm., Autom. Des., 106(2), pp. 156–162. [CrossRef]
Bestle, D. , and Eberhard, P. , 1992, “ Analyzing and Optimizing Multibody Systems,” Mech. Struct. Mach., 20(1), pp. 67–92. [CrossRef]
Eberhard, P. , 1996, “ Adjoint Variable Method for Sensitivity Analysis of Multibody Systems Interpreted as a Continuous, Hybrid Form of Automatic Differentiation,” 2nd International Workshop on Computational Differentiation, Santa Fe, NM, SIAM, Philadelphia, PA, pp. 319–328.
Bottasso, C. , Croce, A. , Ghezzi, L. , and Faure, P. , 2004, “ On the Solution of Inverse Dynamics and Trajectory Optimization Problems for Multibody Systems,” Multibody Syst. Dyn., 11(1), pp. 1–22. [CrossRef]
Bertolazzi, E. , Biral, F. , and Lio, M. D. , 2005, “ Symbolic-Numeric Indirect Method for Solving Optimal Control Problems for Large Multibody Systems,” Multibody Syst. Dyn., 13(2), pp. 233–252. [CrossRef]
Schaffer, A. , 2005, “ On the Adjoint Formulation of Design Sensitivity Analysis of Multibody Dynamics,” Ph.D. thesis, University of Iowa, Iowa City, IA. http://ir.uiowa.edu/cgi/viewcontent.cgi?article=1278&context=etd
Petzold, L. , Li, S. , Cao, Y. , and Serban, R. , 2006, “ Sensitivity Analysis for Differential-Algebraic Equations and Partial Differential Equations,” Comput. Chem. Eng., 30(10–12), pp. 1553–1559. [CrossRef]
Cao, Y. , Li, S. , Petzold, L. , and Serban, R. , 2003, “ Adjoint Sensitivity Analysis for Differential-Algebraic Equations: The Adjoint DAE System and Its Numerical Solution,” SIAM J. Sci. Comput., 24(3), pp. 1076–1089. [CrossRef]
Steiner, W. , and Reichl, S. , 2012, “ The Optimal Control Approach to Dynamical Inverse Problems,” ASME J. Dyn. Syst., Meas., Control, 134(2), p. 021010. [CrossRef]
Nachbagauer, K. , Oberpeilsteiner, S. , Sherif, K. , and Steiner, W. , 2014, “ The Use of the Adjoint Method for Solving Typical Optimization Problems in Multibody Dynamics,” ASME J. Comput. Nonlinear Dyn., 10(6), p. 061011. [CrossRef]
Süli, E. , and Mayers, D. , 2003, An Introduction to Numerical Analysis, Cambridge University Press, Oxford, UK.
Alexe, M. , and Sandu, A. , 2009, “ On the Discrete Adjoints of Adaptive Time Stepping Algorithms,” J. Comput. Appl. Math., 233(4), pp. 1005–1020. [CrossRef]
Negrut, D. , Rampalli, R. , Ottarsson, G. , and Sajdak, A. , 2000, “ On the Use of the HHT Method in the Context of Index 3 Differential Algebraic Equations of Multibody Dynamics,” ASME Paper No. DETC2005-85096.
Hilber, H. , Hughes, T. , and Taylor, R. , 1977, “ Improved Numerical Dissipation for Time Integration Algorithms in Structural Dynamics,” Earthquake Eng. Struct. Dyn., 5(3), pp. 283–292. [CrossRef]
Newmark, N. , 1959, “ A Method of Computation for Structural Dynamics,” J. Eng. Mech. Div., 85(3), pp. 67–94.
Gavrea, B. , Negrut, D. , and Potra, F. A. , 2005, “ The Newmark Integration Method for Simulation of Multibody Systems: Analytical Considerations,” ASME Paper No. IMECE2005-81770.
Nachbagauer, K. , Oberpeilsteiner, S. , and Steiner, S. , 2015, “ Enhancement of the Adjoint Method by Error Control of Accelerations for Parameter Identification in Multibody Dynamics,” Univers. J. Control Autom., 3(3), pp. 47–52. [CrossRef]
Betsch, P. , Quasem, M. , and Uhlar, S. , 2009, “ Numerical Integration of Mechanical Systems With Mixed Holonomic and Control Constraints,” ECCOMAS Thematic Conference on Multibody Dynamics, Warsaw University of Technology, Warszawa, Poland, June 29–July2. http://citeseerx.ist.psu.edu/viewdoc/download?doi=


Grahic Jump Location
Fig. 1

Interpolation of the Hessian

Grahic Jump Location
Fig. 2

Nonlinear two mass oscillator

Grahic Jump Location
Fig. 3

Solution of the nonlinear two mass oscillator (a) optimal input and (b) convergence of cost functional

Grahic Jump Location
Fig. 4

Planar overhead crane (a) schematic of the crane and (b) optimal input

Grahic Jump Location
Fig. 5

Convergence of the two optimization problems (a) convergence of version 1 and (b) convergence of version 2

Grahic Jump Location
Fig. 6

PUMA industrial robot (a) the schematic of the PUMA robot and (b) the start and end configuration

Grahic Jump Location
Fig. 7

Solution of the robot example (a) convergence of the cost functional of the PUMA and (b) optimal joint torques of the PUMA



Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In