Bayesian Optimization for Full Waveform Inversion

  • Bruno G. GaluzziEmail author
  • Riccardo Perego
  • Antonio Candelieri
  • Francesco Archetti
Part of the AIRO Springer Series book series (AIROSS, volume 1)


Full Waveform Inversion (FWI) is a computational method to estimate the physical features of Earth subsurface from seismic data, leading to the minimization of a misfit function between the observed data and the predicted ones, computed by solving the wave equation numerically. This function is usually multimodal, and any gradient-based method would likely get trapped in a local minimum, without a suitable starting point in the basin of attraction of the global minimum. The starting point of the gradient procedure can be provided by an exploratory stage performed by an algorithm incorporating random elements. In this paper, we show that Bayesian Optimization (BO) can offer an effective way to structure this exploration phase. The computational results on a 2D acoustic FWI benchmark problem show that BO can provide a starting point in the parameter space from which the gradient-based method converges to the global optimum.


Bayesian optimization Full waveform inversion Inversion problems 


  1. 1.
    Auer, P.: Using confidence bounds for exploitation-exploration trade-off. J. Mach. Learn. Res. (2002).
  2. 2.
    Beydon, W.B.: First born and rytov approximations: modeling and inversion conditions in a canonical example. J. Acoust. Soc. Am. (1988). Scholar
  3. 3.
    Candelieri, A., Giordani, I., Archetti, F., Barkalov, K., Meyerov, I., Polovinkin, A., Sysoyev, A., Zolotykh, N.: Tuning hyperparameters of a SVM-based water demand forecasting system through parallel global optimization. Comput. Oper. Res. (2018) (Article in Press)Google Scholar
  4. 4.
    Candelieri, A., Perego, R., Archetti, F.: Bayesian optimization of pump operations in water distribution systems. J. Glob. Optim. (Article in Press)Google Scholar
  5. 5.
    Courant, R., Friedrichs, K., Lewy, H.: On the partial difference equations of mathematical physics. IBM J. (1967). Scholar
  6. 6.
    Datta, D., Sen, M.K.: Estimating a starting model for full-waveform inversion. Geophysics (2016). Scholar
  7. 7.
    Fichtner, A.: Full Seismic Waveform Modelling and Inversion. Springer Science & Business Media (2010). Scholar
  8. 8.
    Galuzzi, B., Zampieri, E., Stucchi, E.: A Local adaptive method for the numerical approximation in seismic wave modelling. Commun. Appl. Ind. Math. (2017). Scholar
  9. 9.
    Galuzzi, B., Zampieri, E., Stucchi, E.: Global optimization procedure to estimate a starting velocity model for local Full Waveform Inversion. Optimi. Decis. Sci. Methodol. Appl. (2017). Scholar
  10. 10.
    Mazzotti, A., Bienati, N., Stucchi, E., Tognarelli, A., Aleardi, M., Sajeva, A.: Two-grid genetic algorithm full-waveform inversion. Lead. Edge. (2016). Scholar
  11. 11.
    Močkus, J.: On Bayesian methods for seeking the extremum. In: Optimization Techniques IFIP Technical Conference. Springer (1975). Scholar
  12. 12.
    Močkus, J.: Bayesian Approach to Global Optimization: Theory and Applications. Springer Netherlands (1989). Scholar
  13. 13.
    Morton, K.W., Mayers, D.F.: Numerical Solution of Partial Differential Equations: An Introduction. Cambridge University Press, Cambridge (2005)CrossRefGoogle Scholar
  14. 14.
    Nocedal, J., Wright, S.: Numerical Optimization. Springer (2006).
  15. 15.
    Perdikaris, P., Karniadakis, G.: Model inversion via multi-fidelity Bayesian optimization: a new paradigm for parameter estimation in haemodynamics, and beyond. J. R. Soc. Interface. (2016). Scholar
  16. 16.
    Plessix, R.: A review of the adjoint-state method for computing the gradient of a functional with geophysical applications. Geophys J. Int. (2006). Scholar
  17. 17.
    Sajeva, A., Aleardi, M., Stucchi, E., Bienati, N., Mazzotti, A.: Estimation of acoustic macro models using a genetic full-waveform inversion: applications to the Marmousi model. Geophysics (2016). Scholar
  18. 18.
    The Open University, Walton Hall, Milton Keynes, MK7 \(6AA\) [Online].
  19. 19.
    Versteeg, R.: The Marmousi experience: velocity model determination on a synthetic complex data set. Lead. Edge. (1994). Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  • Bruno G. Galuzzi
    • 1
  • Riccardo Perego
    • 1
  • Antonio Candelieri
    • 1
  • Francesco Archetti
    • 2
  1. 1.Department of Computer Science, Systems and Communications, University of Milano-BicoccaMilanItaly
  2. 2.Consorzio Milano-RicercheMilanItaly

Personalised recommendations