Computational Statistics

, Volume 32, Issue 4, pp 1621–1643 | Cite as

Fast derivatives of likelihood functionals for ODE based models using adjoint-state method

Original Paper
  • 103 Downloads

Abstract

We consider time series data modeled by ordinary differential equations (ODEs), widespread models in physics, chemistry, biology and science in general. The sensitivity analysis of such dynamical systems usually requires calculation of various derivatives with respect to the model parameters. We employ the adjoint state method (ASM) for efficient computation of the first and the second derivatives of likelihood functionals constrained by ODEs with respect to the parameters of the underlying ODE model. Essentially, the gradient can be computed with a cost (measured by model evaluations) that is independent of the number of the ODE model parameters and the Hessian with a linear cost in the number of the parameters instead of the quadratic one. The sensitivity analysis becomes feasible even if the parametric space is high-dimensional. The main contributions are derivation and rigorous analysis of the ASM in the statistical context, when the discrete data are coupled with the continuous ODE model. Further, we present a highly optimized implementation of the results and its benchmarks on a number of problems. The results are directly applicable in (e.g.) maximum-likelihood estimation or Bayesian sampling of ODE based statistical models, allowing for faster, more stable estimation of parameters of the underlying ODE model.

Keywords

Sensitivity analysis Ordinary differential equations Gradient Hessian Statistical computing Mathematical statistics Algorithm 

Notes

Acknowledgements

We would like to thank Xavier Woot de Trixhe from Janssen Pharmaceutica for numerous very interesting discussions on PK/PD, virology, biological pathways modeling, NLMEMs and on life in general. They were an important source of motivation and provided a view from a different perspective. And we would like to thank the anonymous reviewers for their substantial input, enhancing the quality of the paper.

Supplementary material

180_2017_765_MOESM1_ESM.pdf (109 kb)
Supplementary material 1 (pdf 108 KB)
180_2017_765_MOESM2_ESM.r (4 kb)
Supplementary material 2 (R 4 KB)

References

  1. Bazaraa MS, Sherali HD, Shetty CM (2006) Nonlinear programming: theory and algorithms, 3rd edn. John Wiley & Sons, HobokenCrossRefMATHGoogle Scholar
  2. Bensoussan A, Lions J, Papanicolaou G (1978) Asymptotic analysis for periodic structures. North-Holland Pub. Co, AmsterdamMATHGoogle Scholar
  3. Bertsekas DP (1999) Nonlinear programming. Athena scientific, BelmontMATHGoogle Scholar
  4. Brooks S, Gelman A, Jones G, Meng XL (2011) Handbook of markov chain monte carlo. CRC Press, Boca RatonCrossRefMATHGoogle Scholar
  5. Cesari L (1983) Optimization-theory and applications: problems with ordinary differential equations, applications of mathematics, vol 17. Springer-Verlag, New YorkCrossRefMATHGoogle Scholar
  6. Cimrák I, Melicher V (2007) Sensitivity analysis framework for micromagnetism with application to the optimal shape design of magnetic random access memories. Inverse Probl 23(2):563–588CrossRefMATHMathSciNetGoogle Scholar
  7. Coddington EA, Levinson N (1955) Theory of ordinary differential equations. Tata McGraw-Hill Education, New YorkMATHGoogle Scholar
  8. Delyon B, Lavielle M, Moulines E (1999) Convergence of a stochastic approximation version of the EM algorithm. Ann Stat 27(1):94–128CrossRefMATHMathSciNetGoogle Scholar
  9. Draelants D, Broeckhove J, Beemster GTS, Vanroose W (2012) Numerical bifurcation analysis of the pattern formation in a cell based auxin transport model. J Math Biol 67(5):1279–1305CrossRefMATHMathSciNetGoogle Scholar
  10. Gillespie DT (1977) Exact stochastic simulation of coupled chemical reactions. J Phys Chem 81(25):2340–2361CrossRefGoogle Scholar
  11. Haber T, Melicher V, Michiels N, Kovac T, Nemeth B, Claes J (2016) DiffMEM. https://bitbucket.org/tomhaber/diffmem/branch/analysis
  12. Hindmarsh AC, Brown PN, Grant KE, Lee SL, Serban R, Shumaker DE, Woodward CS (2005) SUNDIALS: suite of nonlinear and differential/algebraic equation solvers. ACM Trans Math Softw 31(3):363–396CrossRefMATHMathSciNetGoogle Scholar
  13. Knoll D, Keyes D (2004) Jacobian-free Newton–Krylov methods: a survey of approaches and applications. J Comput Phys 193(2):357–397CrossRefMATHMathSciNetGoogle Scholar
  14. Lavielle M, Samson A, Karina Fermin A, Mentré F (2011) Maximum likelihood estimation of long-term hiv dynamic models and antiviral response. Biometrics 67(1):250–259CrossRefMATHMathSciNetGoogle Scholar
  15. Lewis JM, Lakshmivarahan S, Dhall S (2006) Dynamic data assimilation: a least squares approach, encyclopedia of mathematics and its applications, vol 13. Cambridge University Press, CambridgeCrossRefMATHGoogle Scholar
  16. Lindstrom MJ, Bates DM (1990) Nonlinear mixed effects models for repeated measures data. Biometrics 46(3):673–687CrossRefMathSciNetGoogle Scholar
  17. Lions JL (1971) Optimal control of systems governed by partial differential equations. Springer, BerlinCrossRefMATHGoogle Scholar
  18. Martin J, Wilcox L, Burstedde C, Ghattas O (2012) A stochastic Newton MCMC method for large-scale statistical inverse problems with application to seismic inversion. SIAM J Sci Comput 34(3):A1460–A1487CrossRefMATHMathSciNetGoogle Scholar
  19. Melicher V, Vrábel’ V (2013) On a continuation approach in Tikhonov regularization and its application in piecewise-constant parameter identification. Inverse Probl 29(11):115,008Google Scholar
  20. Moré JJ (1978) The Levenberg–Marquardt algorithm: implementation and theory. In: Watson G (ed) Numerical analysis, lecture notes in mathematics, vol 630, Springer, Berlin, pp 105–116Google Scholar
  21. Murray JD (2002) Mathematical biology I: an introduction, interdisciplinary applied mathematics, vol 17, 3rd edn. Springer-Verlag, New YorkGoogle Scholar
  22. Raue A, Schilling M, Bachmann J, Matteson A, Schelke M, Kaschek D, Hug S, Kreutz C, Harms BD, Theis FJ, Klingmüller U, Timmer J (2013) Lessons learned from quantitative dynamical modeling in systems biology. Plos One 8(9):1–17CrossRefGoogle Scholar
  23. Serban R, Hindmarsh AC (2005) CVODES: the sensitivity-enabled ODE solver in SUNDIALS. In: ASME 2005 international design engineering technical conferences and computers and information in engineering conference, American Society of Mechanical Engineers, pp 257–269Google Scholar
  24. Slodička M, Balážová A (2010) Decomposition method for solving multi-species reactive transport problems coupled with first-order kinetics applicable to a chain with identical reaction rates. Journal of Computational and Applied Mathematics 234(4):1069–1077, proceedings of the Thirteenth International Congress on Computational and Applied Mathematics (ICCAM-2008), Ghent, Belgium, 7–11 July, 2008Google Scholar
  25. Tornøe CW, Agersø H, Jonsson E, Madsen H, Nielsen HA (2004) Non-linear mixed-effects pharmacokinetic/pharmacodynamic modelling in NLME using differential equations. Comput Methods Programs Biomed 76(1):3–40CrossRefGoogle Scholar
  26. Wong R (2001) Asymptotic approximations of integrals, Classics in applied mathematics, vol 34. SIAM, BostonCrossRefGoogle Scholar
  27. Zeidler E (1985) Nonlinear functional analysis and its applications: fixed point theorems, nonlinear functional analysis and its applications. Springer-Verlag, New YorkCrossRefMATHGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany 2017

Authors and Affiliations

  1. 1.Department of Mathematics and Computer ScienceUniversity of AntwerpAntwerpBelgium
  2. 2.Expertise Centre for Digital MediaHasselt UniversityDiepenbeekBelgium

Personalised recommendations