Abstract
Due to increasing popularity of electric vehicles there is rising demand for smart controller solutions that optimize the flow of energy between buildings and electric vehicles. Simple rule-based controllers are (often manually) developed and tuned for specific use case scenarios, for example a specific building, mobility usage patterns and country-specific regulations. However, it is often very difficult to correctly anticipate the exact conditions the controller has to work on so that a high performance under worst-case conditions is a very important target. In this work we use an adversarial optimization approach in order to find both challenging scenarios and controller parameterizations that perform well in those scenarios. We can show that in comparison to a standard controller, our approach can find challenging scenarios for a standard controller and controllers that outperform the baseline on those worst-case scenarios.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Al-Dujaili, A., Srikant, S., Hemberg, E., O’Reilly, U.M.: On the application of Danskin’s theorem to derivative-free minimax optimization. arXiv preprint arXiv:1805.06322 (2018)
Barbosa, H.J.: A coevolutionary genetic algorithm for constrained optimization. In: Proceedings of the 1999 Congress on Evolutionary Computation-CEC99 (Cat. No. 99TH8406), vol. 3, pp. 1605–1611. IEEE (1999)
Ben-Tal, A., El Ghaoui, L., Nemirovski, A.: Robust Optimization, vol. 28. Princeton University Press, Princeton (2009)
Bischof, R., Zedrosser, A.: The educated prey: consequences for exploitation and control. Behav. Ecol. 20(6), 1228–1235 (2009)
Braam, F., Hollinger, R., Engesser, M.L., Müller, S., Kohrs, R., Wittwer, C.: Peak shaving with photovoltaic-battery systems. In: IEEE PES Innovative Smart Grid Technologies, Europe, pp. 1–5. IEEE (2014)
Branke, J.: Evolutionary Optimization in Dynamic Environments, vol. 3. Springer, Heidelberg (2012)
Cheng, R., Rodemann, T., Fischer, M., Olhofer, M., Jin, Y.: Evolutionary many-objective optimization of hybrid electric vehicle control: from general optimization to preference articulation. IEEE Trans. Emerg. Top. Comput. Intell. 1(2), 97–111 (2017)
Cramer, A.M., Sudhoff, S.D., Zivi, E.L.: Evolutionary algorithms for minimax problems in robust design. IEEE Trans. Evol. Comput. 13(2), 444–453 (2008)
ESI-ITI: SimulationX 4.0. http://www.simulationx.com/
Fawzi, A., Fawzi, O., Frossard, P.: Fundamental limits on adversarial robustness. In: Proceedings of ICML, Workshop on Deep Learning (2015)
Halck, O.M., Dahl, F.A.: Asymmetric co-evolution for imperfect-information zero-sum games. In: López de Mántaras, R., Plaza, E. (eds.) ECML 2000. LNCS (LNAI), vol. 1810, pp. 171–182. Springer, Heidelberg (2000). https://doi.org/10.1007/3-540-45164-1_18
Hansen, N., Müller, S., Koumoutsakos, P.: Reducing the time complexity of the derandomized evolution strategy with covariance matrix adaptation. Evol. Comput. 11, 1–18 (2003). http://www.lri.fr/~hansen/publications.html
Herrmann, J.W.: A genetic algorithm for minimax optimization problems. In: Proceedings of the 1999 Congress on Evolutionary Computation-CEC99 (Cat. No. 99TH8406), vol. 2, pp. 1099–1103. IEEE (1999)
Ogino, Y., Iida, R., Rodemann, T.: Using desirability functions for many-objective optimization of a hybrid car controller. In: GECCO 2017 Conference Companion (2017)
Rodemann, T., Kitamura, K.: Simulation-based design and evaluation of a smart energy manager. In: Computer Aided Systems Theory - EUROCAST (2019, to appear)
Sebald, A.V., Schlenzig, J.: Minimax design of neural net controllers for highly uncertain plants. IEEE Trans. Neural Netw. 5(1), 73–82 (1994)
Unger, R., Mikoleit, B., Schwan, T., Bäker, B., Kehrer, C., Rodemann, T.: Green building - modeling renewable building energy systems with emobility using Modelica. In: Proceedings of Modelica 2012 Conference. Modelica Association, Munich, Germany (2012)
Wagner, T., Trautmann, H.: Integration of preferences in hypervolume-based multiobjective evolutionary algorithms by means of desirability functions. IEEE Trans. Evol. Comput. 14(5), 688–701 (2010)
Zhu, L., Deb, K., Kulkarni, S.: Multi-scenario optimization using multi-criterion methods: a case study on byzantine agreement problem. In: 2014 IEEE Congress on Evolutionary Computation (CEC), pp. 2601–2608. IEEE (2014)
Acknowledgments
Baraq Mushtaq acknowledges the financial support from the Honda Research Institute Europe.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Mushtaq, M.B., Rodemann, T. (2020). Adversarial Optimization Approach for Development of Robust Controllers. In: Castillo, P.A., Jiménez Laredo, J.L., Fernández de Vega, F. (eds) Applications of Evolutionary Computation. EvoApplications 2020. Lecture Notes in Computer Science(), vol 12104. Springer, Cham. https://doi.org/10.1007/978-3-030-43722-0_25
Download citation
DOI: https://doi.org/10.1007/978-3-030-43722-0_25
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-43721-3
Online ISBN: 978-3-030-43722-0
eBook Packages: Computer ScienceComputer Science (R0)