Abstract
This paper studies the robust optimal control problem for descriptor systems. We applied differential game theory to solve the disturbance attenuation problem. The robust control problem was converted into a reduced ordinary zero-sum game. Within a linear quadratic setting, we solved the problem for finite and infinite planning horizons.
Similar content being viewed by others
Notes
We would like to thank the referee pointing this out to us.
We write \(X>0 (X\ge 0)\) if \(X\) is positive (semi) definite.
\({\mathbb {C}}^ - = \left\{ {\lambda \in {\mathbb {C}} \left| {{\mathop {\mathrm{Re}}\nolimits } \left( \lambda \right) < 0} \right. } \right\} {\mathbb {C}}_0^ + = \left\{ {\lambda \in {\mathbb {C}}\left| {{\mathop {\mathrm{Re}}\nolimits } \left( \lambda \right) \ge 0} \right. } \right\} .\)
\(\lim _{t_f \rightarrow \infty } J_i \left( {t_f ,x_{1_0 } ,u} \right) = - \infty \left( \infty \right) \) if \(\forall r \in {\mathbb {R}},\exists T_f \in {\mathbb {R}} \text {, such that } t_f \ge T_f\) implies \(J_i \left( {t_f ,x_{1_0 } ,u} \right) \le r\left( { \ge r} \right) \).
Matrix \(A\) is called stable if the real parts of all its eigenvalues are negative.
Such a solution is called an LRS solution.
Note from (2) that the assumption that \(A\) is stable is equivalent to the assumption that all finite eigenvalues of \(A_E\) are stable.
References
Başar, T., Bernhard, P.: \({\cal H}_\infty \) Optimal Control and Related Minimax Design Problem. Modern Birkhäuser Classics, Boston (1995)
van den Broek, W.A., Engwerda, J.C., Schumacher, J.M.: Robust equilibria in indefinite linear-quadratic differential games. JOTA 51, 565–595 (2003)
Kun, G.: Stabilizability, controllability, and optimal strategies of linear and nonlinear dynamical games. Ph.D. Thesis, RWTH-Aachen, Dept. of Mathematics (2001)
Haurie, A., Krawczyk, J., Zaccour, G.: Games and Dynamic Games. World Scientific Publishing Company, Singapore (1993)
Musthofa, M.W., Salmah, Engwerda, J.C., Suparwanto, A.: The open-loop zero-sum linear quadratic impulse free descriptor differential game. Int. J. Appl. Math. Stat. 35(5), 29–44 (2013)
Engwerda, J.C., Salmah, Wijayanti, I.E.: The open-loop discounted linear quadratic differential game for regular higher order index descriptor systems. Int. J. Control 82(12), 2365–2374 (2009)
Pan, Z., Başar, T.: \(H^{-infty}\)-optimal control for singularly perturbed systems. Part I: perfect state measurements. Automatica 29(2), 401–423 (1993)
Pan, Z., Başar, T.: \(H^{-infty}\)-optimal control for singularly perturbed systems. Part II: imperfect measurements. In: Proceedings of 31th IEEE Conference on Decision and Control, Tucson, Arizona (2009)
Houska, B.: Robustness and stability optimization of open-loop controlled power generating kites. Ph.D. Thesis, Ruprecht-Karls-Universität Heidelberg, Fakultät für Mathematik und Informatik (2007)
Gantmacher, F.: Theory of Matrices. Chelsea Publishing Company, New York (1959)
Kautsky, J., Nicholas, N.K., Chu, E.K.-W.: Robust pole assignment in singular control systems. Linear Algebra Appl. 147, 9–37 (1989)
Musthofa, M.W., Salmah, Engwerda, J.C., Suparwanto, A.: The open-loop zero-sum linear quadratic differential game for index one descriptor systems. In: Proceedings of 2nd International Conference on Instrumentation Control and Automation, pp. 350–355 (2011)
Musthofa, M.W.: Robust optimal control design with differential game approach for linear quadratic descriptor systems. Ph.D. thesis, Universitas Gadjah Mada, Dept. of Mathematics (2015)
Engwerda, J.C., Salmah, : The open loop linear quadratic differential game for index one descriptor systems. Automatica 45, 585–592 (2009)
Salmah: Optimal control of regulator descriptor systems for dynamic games. Ph.D. Thesis, Universitas Gadjah Mada, Dept. of Mathematics (2006)
Salmah: Open-loop zero-sum linear quadratic dynamic game for descriptor systems. In: Proceedings of the 1st International Conference on Instrumentation Control and Automation, pp. 179–183 (2009)
Xu, H., Mizukami, K.: Linear-quadratic zero-sum differential games for generalized state space systems. IEEE Trans. Autom. Control 39, 143–147 (1994)
Engwerda, J.C.: Linear Quadratic Dynamic Optimization and Differential Games. Wiley, West Sussex (2005)
Xu, S., Lam, J.: Robust Control and Filtering of Singular Systems. Springer, Berlin (2006)
Wang, H.S., Yung, C.F., Chang, F.R.: \({\cal H}_\infty \) Control for Nonlinear Descriptor System. Springer, London (2006)
Huang, J.C., Wang, H.S., Chang, F.R.: Robust \(H_\infty \) control for uncertain linear time-invariant descriptor systems. Proc. IEE Control Theory Appl. 147, 648–654 (2000)
Lee, H.J., Kau, S.W., Liu, Y.S.: An improvement on robust \({\cal H}_\infty \) control for uncertain continuous-time descriptor systems. Int. J. Control Autom. Syst. 41, 271–280 (2006)
Mattsson, S.E., Söderlind, G.: Index reduction in differential-algebraic equations using dummy derivatives. SIAM J. Sci. Comput. 14, 677–692 (1993)
Kunkel, P., Mehrmann, V.: Index reduction for differential–algebraic equations by minimal extension. Zeitschrift für Angewandte Mathematik and Mechanik 84, 579–597 (2004)
Acknowledgments
The first author is grateful for the support of the Directorate General High Education of Ministry of Education and Culture of Indonesia through a scholarship.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Dean A. Carlson.
Appendix: State Transformation and Shorthand Notation (7, 8)
Appendix: State Transformation and Shorthand Notation (7, 8)
With \(X\) and \(Y\) defined as in (2), \(X =: \left[ {\begin{array}{*{20}c} {X_1 } &{} {X_2 } \\ \end{array}} \right] \) and \(Y^T =: \left[ {\begin{array}{*{20}c} {Y_1^T }&{Y_2^T } \end{array}} \right] \) are nonsingular matrices. With the state transformation \(\left[ {\begin{array}{*{20}c} {x_1 \left( t \right) } \\ {x_2 \left( t \right) } \\ \end{array}} \right] : = X^{ - 1} x\left( t \right) \), the corresponding state and control matrices for the reduced dynamical system are \(A:=J;\ B_1:=Y_1B_u\) and \(B_2:=Y_1B_w\).
The matrices used in Eqs. (7) and (8) are
\(C_1 X =: \bar{C}_1 =: \left[ {\begin{array}{*{20}c} {\bar{C}_{11} }&{\bar{C}_{12} } \end{array}} \right] , C_2 X =: \bar{C}_2 =: \left[ {\begin{array}{*{20}c} {\bar{C}_{21} }&{\bar{C}_{22} } \end{array}} \right] , Q: = X_1^T \bar{Q}X_1 ,\ V: = - X_1^T \bar{Q}X_2 Y_2 B_u ,\ N: = B_u^T Y_2^T X_2^T \bar{Q}X_2 Y_2 B_w ,\ W: = - X_1^T \bar{Q}X_2 Y_2 B_w ,\ R_{\bar{1} \bar{1}} : = B_u^T Y_2^T X_2^T \bar{Q} X_2 Y_2 B_u + \bar{R}_1\), and \(R_{\overline{22}\gamma } : = B_w^T Y_2^T X_2^T \bar{Q}X_2 Y_2 B_w - \gamma \bar{R}_2\).
Rights and permissions
About this article
Cite this article
Musthofa, M.W., Salmah, Engwerda, J. et al. Robust Optimal Control Design Using a Differential Game Approach for Open-Loop Linear Quadratic Descriptor Systems. J Optim Theory Appl 168, 1046–1064 (2016). https://doi.org/10.1007/s10957-015-0750-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10957-015-0750-8
Keywords
- Robust optimal control
- Zero-sum linear quadratic differential game
- Descriptor systems
- Open-loop information structure