Direct methods are becoming the most used technique to solve nonlinear optimal control problems. Regular time meshes having equidistant spacing are frequently used. However, in some cases these meshes cannot cope accurately with nonlinear behavior. One way to improve the solution is to select a new mesh with a greater number of nodes. Another way, involves adaptive mesh refinement. In this case, the mesh nodes have non equidistant spacing which allow a non uniform nodes collocation. In the method presented in this paper, a time mesh refinement strategy based on the local error is developed. After computing a solution in a coarse mesh, the local error is evaluated, which gives information about the subintervals of time domain where refinement is needed. This procedure is repeated until the local error reaches a user-specified threshold. The technique is applied to solve the car-like vehicle problem aiming minimum consumption. The approach developed in this paper leads to results with greater accuracy and yet with lower overall computational time as compared to using a time meshes having equidistant spacing.
Skip Nav Destination
Article navigation
17 October 2013
11TH INTERNATIONAL CONFERENCE OF NUMERICAL ANALYSIS AND APPLIED MATHEMATICS 2013: ICNAAM 2013
21–27 September 2013
Rhodes, Greece
Research Article|
October 17 2013
Mesh refinement strategy for optimal control problems
L. T. Paiva;
L. T. Paiva
Faculty of Engineering, University of Porto, Portugal and Institute of Systems and Robotics, Porto,
Portugal
Search for other works by this author on:
F. A. C. C. Fontes
F. A. C. C. Fontes
Faculty of Engineering, University of Porto, Portugal and Institute of Systems and Robotics, Porto,
Portugal
Search for other works by this author on:
AIP Conf. Proc. 1558, 590–593 (2013)
Citation
L. T. Paiva, F. A. C. C. Fontes; Mesh refinement strategy for optimal control problems. AIP Conf. Proc. 17 October 2013; 1558 (1): 590–593. https://doi.org/10.1063/1.4825560
Download citation file:
Sign in
Don't already have an account? Register
Sign In
You could not be signed in. Please check your credentials and make sure you have an active account and try again.
Sign in via your Institution
Sign in via your InstitutionPay-Per-View Access
$40.00