About DML-CZ | FAQ | Conditions of Use | Math Archives | Contact Us
  Previous |  Up |  Next

Article

Kossaczký, Igor ; Ehrhardt, Mattias ; Günther, Michael
The Tree-Grid Method with Control-Independent Stencil. (English). In: Mikula, Karol (ed.): Proceedings of Equadiff 14. Conference on Differential Equations and Their Applications, Bratislava, July 24-28, 2017. Slovak University of Technology in Bratislava, SPEKTRUM STU Publishing, Bratislava, 2017. pp. 79-88
MSC: 65C40, 65M75
Keywords:
Tree-Grid Method, Hamilton-Jacobi-Bellman equation, Stochastic control problem, Fibonacci algorithm
Summary:
The Tree-Grid method is a novel explicit convergent scheme for solving stochastic control problems or Hamilton-Jacobi-Bellman equations with one space dimension. One of the characteristics of the scheme is that the stencil size is dependent on space, control and possibly also on time. Because of the dependence on the control variable, it is not trivial to solve the optimization problem inside the method. Recently, this optimization part was solved by brute-force testing of all permitted controls. In this paper, we present a simple modification of the Tree-Grid scheme leading to a control-independent stencil. Under such modification an optimal control can be found analytically or with the Fibonacci search algorithm.
References:
[1] Ferguson, D.E.: Fibonaccian searching. Communications of the ACM, 3(12):648, 1960. MR 0115851
[2] Forsyth, P.A., HASH(0x20dc550), Labahn, G.: Numerical methods for controlled Hamilton-Jacobi-Bellman PDEs in finance. Journal of Computational Finance, 11(2):1, 2007. DOI 10.21314/JCF.2007.163
[3] Kilianová, S., HASH(0x20f2568), Ševčovič, D.: A transformation method for solving the Hamilton-Jacobi-Bellman equation for a constrained dynamic stochastic optimal allocation problem. The ANZIAM Journal, 55(01):14–38, 2013. DOI 10.1017/S144618111300031X | MR 3144202
[4] Kossaczký, I., Ehrhardt, M., HASH(0x20f4660), Günther, M.: A new convergent explicit Tree-Grid method for HJB equations in one space dimension. Preprint 17/06, University of Wuppertal, to appear in Numerical Mathematics: Theory, Methods and Applications, 2017. MR 3844119
[5] Kushner, H., HASH(0x20f4f60), Dupuis, P.G.: Numerical methods for stochastic control problems in continuous time. volume 24. Springer Science & Business Media, 2013. MR 1217486
[6] Wang, J., HASH(0x20f7808), Forsyth, P.A.: Maximal use of central differencing for Hamilton-Jacobi-Bellman PDEs in finance. SIAM Journal on Numerical Analysis, 46(3):1580–1601, 2008. DOI 10.1137/060675186 | MR 2391007
[7] Yong, Jiongmin, HASH(0x20f8108), Zhou, Xun Yu: Stochastic controls: Hamiltonian systems and HJB equations. volume 43. Springer Science & Business Media, 1999. MR 1696772
Partner of
EuDML logo