Convergence analysis of a computational method for optimal control of non-linear differential-algebraic systems

1990 ◽  
Vol 21 (11) ◽  
pp. 2337-2350 ◽  
Author(s):  
CHYI-TSONG CHEN ◽  
CHYI HWANG
2019 ◽  
Vol 23 (Suppl. 1) ◽  
pp. 275-283
Author(s):  
Kubra Bicer ◽  
Mehmet Sezer

In this paper, a matrix method is developed to solve quadratic non-linear differential equations. It is assumed that the approximate solutions of main problem which we handle primarily, is in terms of Bernoulli polynomials. Both the approximate solution and the main problem are written in matrix form to obtain the solution. The absolute errors are applied to numeric examples to demonstrate efficiency and accuracy of this technique. The obtained tables and figures in the numeric examples show that this method is very sufficient and reliable for solution of non-linear equations. Also, a formula is utilized based on residual functions and mean value theorem to seek error bounds.


Automatica ◽  
2006 ◽  
Vol 42 (6) ◽  
pp. 959-965 ◽  
Author(s):  
Jan Åslund ◽  
Erik Frisk

2011 ◽  
Vol Volume 14 - 2011 - Special... ◽  
Author(s):  
F. Nahayo ◽  
S. Khardi ◽  
J. Ndimubandi ◽  
M. Haddou ◽  
M. Hamadiche

International audience This contribution aims to develop an acoustic optimization model of flight paths minimizing two-aircraft perceived noise on the ground. It is about minimizing the noise taking into account all the constraints of flight without conflict. The flight dynamics associated with a cost function generate a non-linear optimal control problem governed by ordinary non-linear differential equations. To solve this problem, the theory of necessary conditions for optimal control problems with instantaneous constraints is well used. This characterizes the optimal solution as a local one when the newtonian approach has been used alongside the optimality conditions of Karush-Kuhn-Tucker and the trust region sequential quadratic programming. The SQP methods are suggested as an option by commercial KNITRO solver under AMPL programming language. Among several possible solution, it was shown that there is an optimal trajectory (for each aircraft) leading to a reduction of noise levels on the ground. Cette contribution vise à développer un modèle mathématique d’optimisation acoustique des trajectoires de vol de deux avions en approche et sans conflit, en minimisant le bruit perçu au sol. Toutes les contraintes de vol des deux avions sont considérées. La dynamique de vol associée au coût génère un problème de contrôle optimal régis par des équations différentielles ordinaires non-linéaires. Pour résoudre ce problème, la théorie des conditions nécessaires d’optimalité pour des problèmes de commande optimale avec contraintes instanées est bien développée. Ceci se caractérise par une solution optimale locale lorsque l’approche newtonienne est utilisée en tenant compte des conditions d’optimalité de Karush-Kuhn-Tucker et la programmation quadratique séquentielle globalisée par région de confiance. Les méthodes SQP sont proposées comme option par KNITRO sous le langage de programmation AMPL. Parmi plusieurs solutions admissibles, il est retenu une trajectoire optimale menant à une réduction du niveau de bruit au sol.


2021 ◽  
Vol 153 ◽  
pp. 108041
Author(s):  
Lakshay Jain ◽  
Mohanakrishnan Prabhakaran ◽  
Ramamoorthy Karthikeyan ◽  
Umasankari Kannan

1980 ◽  
Vol 25 (92) ◽  
pp. 229-246 ◽  
Author(s):  
L. W. Morland ◽  
I. R. Johnson

AbstractSteady plane flow under gravity of a symmetric ice sheet resting on a horizontal rigid bed, subject to surface accumulation and ablation, basal drainage, and basal sliding according to a shear-traction-velocity power law, is treated. The surface accumulation is taken to depend on height, and the drainage and sliding coefficient also depend on the height of overlying ice. The ice is described as a general non-linearly viscous incompressible fluid, with illustrations presented for Glen’s power law, the polynomial law of Colbeck and Evans, and a Newtonian fluid. Uniform temperature is assumed so that effects of a realistic temperature distribution on the ice response are not taken into account. In dimensionless variables a small paramter ν occurs, but the ν = 0 solution corresponds to an unbounded sheet of uniform depth. To obtain a bounded sheet, a horizontal coordinate scaling by a small factor ε(ν) is required, so that the aspect ratio ε of a steady ice sheet is determined by the ice properties, accumulation magnitude, and the magnitude of the central thickness. A perturbation expansion in ε gives simple leading-order terms for the stress and velocity components, and generates a first order non-linear differential equation for the free-surface slope, which is then integrated to determine the profile. The non-linear differential equation can be solved explicitly for a linear sliding law in the Newtonian case. For the general law it is shown that the leading-order approximation is valid both at the margin and in the central zone provided that the power and coefficient in the sliding law satisfy certain restrictions.


Sign in / Sign up

Export Citation Format

Share Document