Marine Ecosystem Model Calibration through Enhanced Surrogate-Based Optimization - Simulation and Modeling Methodologies, Technologies and Application - page 198

such as surrogate-based optimization (SBO) techniques, are highly desirable. The idea
of SBO is to replace the high-fidelity model in the optimization run by a surrogate, its
computationally cheap and yet reasonably accurate representation.
As a case study, we have investigated parameter optimization of a representative of
the class of one-dimensional marine ecosystem models. As demonstrated in our previ-
ous work, a simple multiplicative response correction applied to a temporally coarser
discretized physics-based low-fidelity (coarse) model of the system of interest is suffi-
cient to create a reliable surrogate of the original, high-fidelity ecosystem model, which
can be used as a prediction tool to calibrate the latter. This approach allowed us to yield
remarkably good results, both in terms of the quality of the final solution and, most
importantly, in terms of the relative reduction in the total optimization cost, about 84%
when compared to the direct fine model optimization.
In this paper, we demonstrated that the correction scheme can be enhanced to alle-
viate the difficulties of its original version, which results in further improvement of the
surrogate model accuracy and overall performance of the optimization algorithm utiliz-
ing this surrogate. The optimization cost was reduced by a factor of three (from 16% to
5% of the direct high-fidelity model optimization optimization cost), which corresponds
to the cost savings of 95%.
Improvements of the present approach by utilizing additionally sensitivity informa-
tion of the low- and the high-fidelity model in the alignment of the low-fidelity model
as well as trust-region convergence safeguards applied to enhance the optimization pro-
cess are expected to further improve the robustness of the algorithm and the accuracy
of the solution. The trade-offs between the accuracy and extra costs due too sensitivity
evaluation will have to be inspected.
Acknowledgements. The authors would like to thank Andreas Oschlies, IFM Geomar,
Kiel. This research was supported by the DFG Cluster of Excellence Future Ocean.
References
1. Bandler, J.W., Cheng, Q.S., Dakroury, S.A., Mohamed, A.S., Bakr, M.H., Madsen, K.,
Søndergaard, J.: Space mapping: The state of the art. IEEE T. Microw. Theory 52(1) (2004)
2. Banks, H.T., Kunisch, K.: Estimation Techniques for Distributed Parameter Systems.
Birkhauser (1989)
3. Bucker, H.M., Fortmeier, O., Petera, M.: Solving a parameter estimation problem in a three-
dimensional conical tube on a parallel and distributed software infrastructure. Journal of
Computational Science 2(2), 95-104 (2011); Simulation Software for Supercomputers
4. Conn, A.R., Gould, N.I.M., Toint, P.L.: Trust-region methods. Society for Industrial and
Applied Mathematics, Philadelphia (2000)
5. Fennel, W., Neumann, T.: Introduction to the Modelling of Marine Ecosystems. Elsevier
(2004)
6. Forrester, A.I.J., Keane, A.J.: Recent advances in surrogate-based optimization. Prog.
Aerosp. Sci. 45(1-3), 50-79 (2009)
7. Gill, A.E.: Atmosphere - Ocean Dynamics. International Geophysics Series, vol. 30. Aca-
demic Press (1982)
8. Koziel, S., Bandler, J.W., Cheng, Q.S.: Robust trust-region space-mapping algorithms for
microwave design optimization. IEEE T. Microw. Theory 58(8), 2166-2174 (2010)