A max-plus based fundamental solution to a class of linear regulator problems with non-quadratic terminal payoff

This paper studies a class of linear regulator problem where the terminal payoff function is not necessarily quadratic. The value function for this problem is generally not quadratic and thus it can not be reduced to solving the corresponding matrix Riccati equation as for the standard linear quadratic regulator (LQR) problem. The computational method of direct iteration using the dynamic programming equations is computationally expensive. In this paper, a new computational method based on max-plus techniques is developed for this problem which is demonstrated to be more efficient and more accurate. In particular, three max-plus fundamental solutions are obtained which can be used as the kernel of max-plus integration with respect to the max-plus dual of the terminal payoff to generate the value function of the linear regulator problem.

[1]  William M. McEneaney,et al.  Max-plus methods for nonlinear control and estimation , 2005 .

[2]  J. Quadrat,et al.  Max-Plus Algebra and System Theory: Where We Are and Where to Go Now , 1999 .

[3]  William M. McEneaney,et al.  Max-plus eigenvector representations for solution of nonlinear H∞ problems: basic concepts , 2003, IEEE Trans. Autom. Control..

[4]  B. Anderson,et al.  Optimal control: linear quadratic methods , 1990 .

[5]  William M. McEneaney,et al.  A Curse-of-Dimensionality-Free Numerical Method for Solution of Certain HJB PDEs , 2007, SIAM J. Control. Optim..

[6]  Geert Jan Olsder,et al.  Synchronization and Linearity: An Algebra for Discrete Event Systems , 1994 .

[7]  Stéphane Gaubert,et al.  The Max-Plus Finite Element Method for Solving Deterministic Optimal Control Problems: Basic Properties and Convergence Analysis , 2008, SIAM J. Control. Optim..

[8]  D. Mayne,et al.  On the discrete time matrix Riccati equation of optimal control , 1970 .

[9]  William M. McEneaney,et al.  A new fundamental solution for differential Riccati equations arising in control , 2008, Autom..

[10]  William M. McEneaney,et al.  A max-plus method for optimal control of a diffusion equation , 2012, 2012 IEEE 51st IEEE Conference on Decision and Control (CDC).

[11]  William M. McEneaney,et al.  Max-Plus Eigenvector Methods for Nonlinear Hinfinity Problems: Error Analysis , 2004, SIAM J. Control. Optim..

[12]  William M. McEneaney,et al.  A max-plus based fundamental solution for a class of infinite dimensional Riccati equations , 2011, IEEE Conference on Decision and Control and European Control Conference.