High order structure preserving explicit methods for solving linear-quadratic optimal control problems

[EN] We consider the numerical integration of linear-quadratic optimal control problems. This problem requires the solution of a boundary value problem: a non-autonomous matrix Riccati differential equation (RDE) with final conditions coupled with the state vector equation with initial conditions. The RDE has positive definite matrix solution and to numerically preserve this qualitative property we propose first to integrate this equation backward in time with a sufficiently accurate scheme. Then, this problem turns into an initial value problem, and we analyse splitting and Magnus integrators for the forward time integration which preserve the positive definite matrix solutions for the RDE. Duplicating the time as two new coordinates and using appropriate splitting methods, high order methods preserving the desired property can be obtained. The schemes make sequential computations and do not require the storrage of intermediate results, so the storage requirements are minimal. The proposed methods are also adapted for solving linear-quadratic N-player differential games. The performance of the splitting methods can be considerably improved if the system is a perturbation of an exactly solvable problem and the system is properly split. Some numerical examples illustrate the performance of the proposed methods.The author wishes to thank the University of California San Diego for its hospitality where part of this work was done. He also acknowledges the support of the Ministerio de Ciencia e Innovacion (Spain) under the coordinated project MTM2010-18246-C03. The author also acknowledges the suggestions by the referees to improve the presentation of this work.Blanes Zamora, S. (2015). High order structure preserving explicit methods for solving linear-quadratic optimal control problems. Numerical Algorithms. 69:271-290. https://doi.org/10.1007/s11075-014-9894-027129069Abou-Kandil, H., Freiling, G., Ionescy, V., Jank, G.: Matrix Riccati equations in control and systems theory. Basel, Burkhäuser Verlag (2003)Al-Mohy, A.H., Higham, N.J.: Computing the Action of the Matrix Exponential, with an Application to Exponential Integrators. SIAM. J. Sci. Comp. 33, 488–511 (2011)Anderson, B.D.O., Moore, J.B.: Optimal control: linear quadratic methods. Dover, New York (1990)Ascher, U.M., Mattheij, R.M., Russell, R.D.: Numerical solutions of boundary value problems for ordinary differential equations. Prentice-Hall, Englewood Cliffs (1988)Bader, P., Blanes, S., Ponsoda, E.: Structure preserving integrators for solving linear quadratic optimal control problems with applications to describe the flight of a quadrotor. J. Comput. Appl. Math. 262, 223–233 (2014)Basar, T., Olsder, G.J.: Dynamic non cooperative game theory, 2nd Ed, SIAM, Philadelphhia (1999)Blanes, S., Casas, F.: On the necessity of negative coefficients for operator splitting schemes of order higher than two. Appl. Num. Math. 54, 23–37 (2005)Blanes, S., Casas, F., Farrés, A., Laskar, J., Makazaga, J., Murua, A.: New families of symplectic splitting methods for numerical integration in dynamical astronomy. Appl. Numer. Math. 68, 58–72 (2013)Blanes, S., Casas, F., Oteo, J.A., Ros, J.: The Magnus expansion and some of its applications. Phys. Rep. 470, 151–238 (2009)Blanes, S., Casas, F., Ros, J.: High order optimized geometric integrators for linear differential equations. BIT 42, 262–284 (2002)Blanes, S., Diele, F., Marangi, C., Ragni, S.: Splitting and composition methods for explicit time dependence in separable dynamical systems. J. Comput. Appl. Math. 235, 646–659 (2010)Blanes, S., Moan, P.C.: Practical symplectic partitioned Runge-Kutta and Runge-Kutta-Nystrm methods. J. Comput. Appl. Math. 142, 313–330 (2002)Blanes, S., Ponsoda, E.: Magnus integrators for solving linear-quadratic differential games. J. Comput. Appl. Math. 236, 3394–3408 (2012)Brif, C., Chakrabarti, R., Rabitz, H.: Control of quantum phenomena: past, present and future. New J. Phys. 12, 075008(68pp) (2010)Cruz, J.B., Chen, C.I.: Series Nash solution of two person non zero sum linear quadratic games. J. Optim. Theory Appl. 7, 240–257 (1971)Dieci, L., Eirola, T.: Positive definitness in the numerical solution of Riccati differential quations. Numer. Math. 67, 303–313 (1994)Engwerda, J.: LQ dynamic optimization and differential games. Wiley (2005)Hairer, E., Lubich, C., Wanner, G.: Geometric Numerical Integration. Structure-Preserving Algorithms for Ordinary Differential Equations (2nd edition). Springer Series in Computational Mathematics, 31. Springer-Verlag (2006)Hochbruck, M., Ostermann, A.: Exponential integrators. Acta Numerica 19, 209–286 (2010)Horn, R.A., Johnson, C.R.: Matrix Analysis. Cambridge University Press, New York (1985)Iserles, A., Munthe-Kaas, H.Z., Nørsett, S.P., Zanna, A.: Lie group methods. Acta Numerica 9, 215–365 (2000)Iserles, A., Nørsett, S.P.: On the solution of linear differential equations in Lie groups. Phil. Trans. R. Soc. Lond. A 357, 983–1019 (1999)Jódar, L., Ponsoda, E.: Non-autonomous Riccati-type matrix differential equations: existence interval, construction of continuous numerical solutions and error bounds. IMA. J. Num. Anal. 15, 61–74 (1995)Jódar, L., Ponsoda, E., Company, R.: Solutions of coupled Riccati equations arising in differential games. Control. Cybern. 24, 117–128 (1995)Kaitala, V, Pohjola, M. In: Carraro, Filar (eds.) : Sustainable international agreement on greenhouse warming. A game theory study. Control and Game Theoretic Models of the Environment, pp 67–87. Birkhauser, Boston (1995)Keller, H.B.: Numerical solution of two point boundary value problems. In: CBMS-NSF Regional Conference Series in Applied Mathematics, Vol. 24. SIAM, Philadelphia (1976)McLachlan, R.I.: Composition methods in the presence of small parameters. BIT 35, 258–268 (1995)McLachlan, R.I., Quispel, R.: Splitting Methods. Acta Numer. 11, 341–434 (2002)Moler, C.B., Van Loan, C.F.: Nineteen Dubious Ways to Compute the Exponential of a Matrix, twenty-five years later. SIAM Rev. 45, 3–49 (2003)Na, T.Y.: Computational methods in engineering boundary value problems. In: Mathematics in Science and Engineering, Vol. 145. Accademic Press, New York (1979)Palao, J.P., Kosloff, R.: Quantum computing by an optimal control algorithm for unitry transformations. Phys. Rev. Lett. 28 (2002)Peirce, A.P., Dahleh, M.A., Rabitz, H.: Optimal control of quantum-mechanical systems: existence, numerical approximation, and applications. Phys. Rev. A 37, 4950–4967 (1988)Reid, W.T.: Riccati Differential Equations. Academic, New York (1972)Sanz-Serna, J.M., Calvo, M.P.: Numerical Hamiltonian Problems. Chapman & Hall, London (1994)Sidje, R.B.: Expokit: a software package for computing matrix exponentials. ACM Trans. Math. Software 24, 130–156 (1998)Speyer, J.L., Jacobson, D.H.: Primer on optimal control theory. SIAM, Philadelphia (2010)Starr, A.W., Ho, Y.C.: Non-zero sum differential games. J. Optim. Theory and Appl 3, 179–197 (1969)Zhu, W., Rabitz, H.: A rapid monotonically convergent iteration algorithm for quantum optimal control ever the expectation value of a positive definite operator. J. Chem. Phys. 109, 385–391 (1998