Creating and sharing knowledge for telecommunications

Dynamic Canonical Descent and Densification: A Direct Method for Optimal Control

Bousson, K. B. ; Correia, S. D. Correia

Dynamic Canonical Descent and Densification: A Direct Method for Optimal Control, Proc IASTED International Conf. on Modelling, Identification and Control - MIC, Innsbruck, Austria, Vol. 1, pp. 1 - 5, February, 2005.

Digital Object Identifier:

Abstract
The present paper proposes an optimization algorithm for dealing with optimal control issues based on direct methods. Since direct methods involve the discretization of the differential equations modelling the dynamics of the system and the cost function, the dimension of the resulting problem may be very high, which can be a challenge for most optimization algorithms. In this paper, the Dynamic Canonical Decent Algorithm is used with the densification of the search domain to reduce the number of variables of the original cost function. A novel approach of domain densification is proposed, along with a method for reducing the dimension of the search domain.