Simple Search

Query:

Enter a query and click »Search«...

Format:

Display: entries per page entries

Zbl 1072.49004
Cabot, Alexandre
The steepest descent dynamical system with control. Applications to constrained minimization. (English)
[J] ESAIM, Control Optim. Calc. Var. 10, 243-258 (2004). ISSN 1292-8119; ISSN 1262-3377/e

Summary: Let $H$ be a real Hilbert space, $\Phi_1:H\to\bbfR$ a convex function of class ${\cal C}^1$ that we wish to minimize under the convex constraint $S$. A classical approach consists in following the trajectories of the generalized steepest descent system (cf. [{\it H Brézis}, ``Opérateurs maximaux monotones et semi-groupes de contractions dans les espaces de Hilbert'' (1999; Zbl 0252.47055)]) applied to the non-smooth function $\Phi_1+\delta_S$. Following {\it A. S. Antipin} [Differ. Equations 30, No. 9, 1365--1375 (1994); translation from Differ. Uravn. 30, No. 9, 1475--1486 (1994; Zbl 0852.49021)] it is also possible to use a continuous gradient-projection system. We propose here an alternative method as follows: given a smooth convex function $\Phi_0:H\to \bbfR$ whose critical points coincide with $S$ and a control parameter $\varepsilon:\bbfR_+\to\bbfR_+$ tending to zero, we consider the ``Steepest Descent and Control'' system $$\dot x(t)+\nabla \Phi_0 \bigl(x(t)\bigr)+ \varepsilon(t)\nabla\Phi_1\bigl(x(t)\bigr)=0,\tag SDC$$ where the control $\varepsilon$ satisfies $\int_0^{+\infty} \varepsilon (t)dt=+\infty$. This last condition ensures that $\varepsilon$ ``slowly" tends to 0. When $H$ is finite dimensional, we then prove that $d(x(t), \text{argmin}_S\Phi_1)\to 0$ $(t\to+ \infty)$, and we give sufficient conditions under which $x(t)\to\overline x\in \text {argmin}_S \Phi_1$. We end the paper by numerical experiments allowing to compare the (SDC) system with the other systems already mentioned.

MSC 2000:: *49J24 Optimal control problems with differential inclusions (existence)
34D05 Asymptotic stability of ODE
34G20 Nonlinear ODE in abstract spaces
37N40 Dynamical systems in optimization and economics
90C48 Programming in abstract spaces

Keywords: dissipative dynamical system; steepest descent method; constrained optimization; convex minimization; asymptotic behaviour; non-linear oscillator

Citations: Zbl 0252.47055; Zbl 0852.49021

Comment on this Item PDF MathML XML ASCII DVI PS BibTeX Online Ordering Article Article Article Journal Journal

	Scientific prize winners of the ICM 2010
	Overhang
	Lie groups, physics and geometry. An introduction for physicists, engineers and chemists.

Simple Search

Zentralblatt MATH Berlin [Germany]