Abstract
We propose a new method for solving nonlinear dynamic tracking games using a meta-heuristic approach. In contrast to ‘traditional’ methods based on linear-quadratic (LQ) techniques, this derivative-free method is very flexible with regard to the objective function specification. The proposed method is applied to a three-player dynamic game and tested versus a derivative-dependent method in approximating solutions of different game specifications. In particular, we consider a dynamic game between fiscal (played by national governments) and monetary policy (played by a central bank) in a monetary union. Apart from replicating results of the LQ-based techniques in a standard setting, we solve two ‘non-standard’ extensions of this game (dealing with an inequality constraint in a control variable and introducing asymmetry in penalties of the objective function), identifying both a cooperative Pareto and a non-cooperative open-loop Nash equilibria, where the traditional methods are not applicable. We, thus, demonstrate that the proposed method allows one to study more realistic problems and gain better insights for economic policy.
Acknowledgements
This work has benefited from a presentation at the WEHIA workshop in Nice. IS acknowledges support from the German Science Foundation (DFG RTG 1411), the Helmholtz Association (HIRG-0069), Projex CSES, Initiative d’Excellence, Université de Strasbourg and RFBR grant Nr. 18-010-01190.
References
Assenmacher-Wesche, K. (2006), Estimating Central Banks’ Preferences from a Time-varying Empirical Reaction Function. European Economic Review 50 (8): 1951–1974.10.1016/j.euroecorev.2005.10.003Search in Google Scholar
Başar, T., Olsder G.J. (1998), Dynamic Noncooperative Game Theory. Philadelphia, SIAM.10.1137/1.9781611971132Search in Google Scholar
Bednar-Friedl, B., Behrens D., Getzner M. (2012), Optimal Dynamic Control of Visitors and Endangered Species in a National Park. Environmental and Resource Economics 52 (1): 1–22.10.1007/s10640-011-9515-5Search in Google Scholar
Bellman, R. (1957), Dynamic Programming. Princeton University Press, Princeton.Search in Google Scholar
Blueschke, D., Neck R. (2011), “Core” and “Periphery” in a Monetary Union: A Macroeconomic Policy Game. International Advances in Economic Research 17: 334–346.10.1007/s11294-011-9303-6Search in Google Scholar
Blueschke, D., Savin I. (2017), No Such Thing Like Perfect Hammer: Comparing Different Objective Function Specifications for Optimal Control. Central European Journal of Operations Research 25: 377–392.10.1007/s10100-016-0446-7Search in Google Scholar
Blueschke, D., Blueschke-Nikolaeva V., Savin I. (2013a), New Insights into Optimal Control of Nonlinear Dynamic Econometric Models: Application of a Heuristic Approach. Journal of Economic Dynamics and Control 37 (4): 821–837.10.1016/j.jedc.2012.12.003Search in Google Scholar
Blueschke, D., Neck R., Behrens D.A. (2013b), OPTGAME3: A Dynamic Game Solver and an Economic Example. 29–51 in: V. Krivan, G. Zaccour (eds.), Advances in Games Dynamic. Theory, Applications, and Numerical Methods, Birkhäuser Verlag. Basel, Switzerland.10.1007/978-3-319-02690-9_2Search in Google Scholar
Boryczka, U., Juszczuk P. (2013), Differential Evolution as a New Method of Computing Nash Equilibria. in: Nguyen Ngoc Thanh (ed.), Transactions on Computational Collective Intelligence IX. Lecture Notes in Science Computer, vol 7770, Springer. Craiova; Romania.10.1007/978-3-642-36815-8_9Search in Google Scholar
Brumm, J., Scheidegger S. (2017), Using Adaptive Sparse Grids to Solve High-dimensional Dynamic Models. Econometrica 85 (5): 1575–1612.10.3982/ECTA12216Search in Google Scholar
Cruz, I., Willigenburg L., Straten G. (2003), Efficient Differential Evolution Algorithms for Multimodal Optimal Control Problems. Applied Soft Computing 3: 97–122.10.1016/S1568-4946(03)00007-3Search in Google Scholar
Cukierman, A. (2002), Are Contemporary Central Banks Transparent about Economic Models and Objectives and What Difference Does it Make? The Federal Reserve Bank of St Louis, pp. 15–35.Search in Google Scholar
Das, S., Suganthan N. (2011), Differential Evolution: A Survey of the State-of-the-art. IEEE Transactions on Evolutionary Computation 15 (1): 4–31.10.1109/TEVC.2010.2059031Search in Google Scholar
Dennis, R. (2004), Inferring Policy Objectives from Economic Outcomes. Oxford Bulletin of Economics and Statistics 66: 735–764.10.1111/j.1468-0084.2004.100_1.xSearch in Google Scholar
Dockner, E.J., Long N.V. (1993), International Pollution Control: Cooperative versus Noncooperative Strategies. Journal of Environmental Economics and Management 25 (1): 13–29.10.1006/jeem.1993.1023Search in Google Scholar
Eggertsson, G., Woodford M. (2003), Zero Bound on Interest Rates and Optimal Monetary Policy. Brookings Papers on Economic Activity, pp. 139–233.10.1353/eca.2003.0010Search in Google Scholar
Engwerda, J., van Aarle B., Plasmans J. (2002), Cooperative and Non-cooperative Fiscal Stabilization Policies in the EMU. Journal of Economic Dynamics and Control 26 (3): 451–481.10.1016/S0165-1889(00)00056-7Search in Google Scholar
Friedman, B.M. (1972), Optimal Economic Stabilization Policy: An Extended Framework. Journal of Political Economy 80: 1002–1022.10.1086/259947Search in Google Scholar
Gilli, M., Schumann E. (2011), Optimal Enough? Journal of Heuristics 17 (4): 373–387.10.1007/s10732-010-9138-ySearch in Google Scholar
Gilli, M., Schumann E. (2014), Optimization Cultures. WIREs Computational Statistics 6 (5): 352–358.10.1002/wics.1312Search in Google Scholar
Gilli, M., Winker P. (2009), Heuristic Optimization Methods in Econometrics. 81–119 in: D. Belsley, E. Kontoghiorghes (eds.), Handbook of Computational Econometrics. Wiley, Chichester.10.1002/9780470748916.ch3Search in Google Scholar
Herrmann, J., Savin I. (2017), Optimal Policy Identification: Insights from the German Electricity Market. Technological Forecasting and Social Change 122: 71–90.10.1016/j.techfore.2017.04.014Search in Google Scholar
Krause, MU., Moyen S. (2016), Public Debt and Changing Inflation Targets. American Economic Journal: Macroeconomics 8 (4): 142–176.10.1257/mac.20130014Search in Google Scholar
Leitmann, G. (1974), Cooperative and Non-Cooperative Many Player Differential Games. Springer, Vienna.10.1007/978-3-7091-2914-2Search in Google Scholar
Leon, M., Xiong N. (2016), Adapting Differential Evolution Algorithms for Continuous Optimization via Greedy Adjustment of Control Parameters. Journal of Artificial Intelligence and Soft Computing Research 6 (2): 103–118.10.1515/jaiscr-2016-0009Search in Google Scholar
Mandler, M. (2012), Decomposing Federal Funds Rate Forecast Uncertainty Using Time-varying Taylor Rules and Real-time Data. The North American Journal of Economics and Finance 23 (2): 228–245.10.1016/j.najef.2012.01.003Search in Google Scholar
Modares, H., Sistani M. (2011), Solving Nonlinear Optimal Control Problems Using a Hybrid IPSO-SQP Algorithm. Engineering Applications of Artificial Intelligence 24: 476–484.10.1016/j.engappai.2010.08.002Search in Google Scholar
Nakov, A. (2008), Optimal and Simple Monetary Policy Rules with Zero Floor on the Nominal Interest Rate. International Journal of Central Banking 4 (2): 73–127.Search in Google Scholar
Neck, R., Blueschke D. (2014), “Haircuts” for the EMU Periphery: Virtue or Vice? Empirica 41 (2): 153–175.Search in Google Scholar
Nobay, A.R., Peel D.A. (2003), Optimal Discretionary Monetary Policy in a Model of Asymmetric Central Bank Preferences. Economic Journal 113: 657–665.10.1111/1468-0297.t01-1-00149Search in Google Scholar
Oliveira, H. (2016), Nash Equilibria of Finite Strategic Games and Fuzzy ASA. 77–92. in: Optimization Evolutionary Global, Manifolds and Applications, Studies in Systems, Decision and Control, Heidelberg vol 43.10.1007/978-3-319-26467-7_5Search in Google Scholar
Oliveira, H., Petraglia A. (2014), Establishing Nash Equilibria of Strategic Games: A Multistart Fuzzy Adaptive Simulated Annealing Approach. Applied Soft Computing 19: 188–197.10.1016/j.asoc.2014.02.013Search in Google Scholar
Pontryagin, L., Boltyanskii V., Gamkrelidze R.V., Mishchenko E.F. (1962), The Mathematical Theory of Optimal Processes. English Translation: Interscience, New York.Search in Google Scholar
Savin, I., Blueschke D. (2016), Lost in Translation: Explicitly Solving Nonlinear Stochastic Optimal Control Problems Using the Median Objective Value. Computational Economics 48 (2): 317–338.10.1007/s10614-015-9526-3Search in Google Scholar
Storn, R., Price K. (1997), Differential Evolution: A Simple and Efficient Adaptive Scheme for Global Optimization Over Continuous Spaces. Journal of Global Optimization 11: 341–359.10.1023/A:1008202821328Search in Google Scholar
Varian, H. (1974), A Bayesian Approach to Real Estate Assessment. in: Feinberg S., Zellner A. (eds.), Studies in Bayesian Economics in Honour of L.J. Savage, North-Holland, Amsterdam.Search in Google Scholar
Zellner, A. (1986), Bayesian Estimation and Prediction Using Asymmetric Loss Functions. Journal of the American Statistical Association 81: 446–451.10.1080/01621459.1986.10478289Search in Google Scholar
Zhuang, Y., Huang H. (2014), Time-optimal Trajectory Planning for Underactuated Spacecraft Using a Hybrid Particle Swarm Optimization Algorithm. Acta Astronautica 94 (2): 690–698.10.1016/j.actaastro.2013.06.023Search in Google Scholar
Appendix
As a stochastic optimization algorithm DE does not claim to identify “the global optimum” but just a good approximation of it. This is because it only asymptotically converges towards the global optimum using the allocated number of restarts and iterations. We report our results based on ten restarts (Table 6) and demonstrate very small differences between them. However, similar values of the aggregated objective value do not necessarily result from the same equilibrium (i.e. similar approximations of it). Figure 6 provides graphical illustration of the optimal paths from different restarts for the control variables. We concentrate on the control variables only as those define the outcome of the dynamic system. Our results unequivocally indicate that we get similar approximations of the same equilibrium. In addition, in Table 8 the Euclidian distance for the normalized (subtracted mean value and divided by standard deviation) control variable paths is reported for each restart from the “best” restart (the row of the best restart is then 0).
Euclidian distances between the normalized values of control variables resulting from 10 restarts and the best solution.
| #restart | Pareto | Nash |
| 1 | 0.53 | 0 |
| 2 | 0.17 | 1.19 |
| 3 | 0.16 | 1.10 |
| 4 | 0.25 | 1.17 |
| 5 | 0.86 | 1.45 |
| 6 | 0 | 1.13 |
| 7 | 0.89 | 1.00 |
| 8 | 0.27 | 1.07 |
| 9 | 0.42 | 1.23 |
| 10 | 0.23 | 1.21 |
Optimal paths from 10 restarts for the control variables
Code and Datasets
The author(s) published code and data associated with this article in the ZBW Journal Data Archive, a storage platform for datasets. See: https://doi.org/10.15456/jbnst.2018075.092731.
© 2018 Oldenbourg Wissenschaftsverlag GmbH, Published by De Gruyter Oldenbourg, Berlin/Boston
