# American Institute of Mathematical Sciences

ISSN:
1078-0947

eISSN:
1553-5231

All Issues

## Discrete & Continuous Dynamical Systems - A

July 1995 , Volume 1 , Issue 3

Select all articles

Export/Reference:

1995, 1(3): 303-346 doi: 10.3934/dcds.1995.1.303 +[Abstract](1132) +[PDF](373.8KB)
Abstract:
We will consider the family of wave equations with boundary damping

$\qquad\qquad \qquad\qquad \epsilon u_{t t} -\Delta u + \lambda u =f$ on $\Omega \times (0,T)$

$(P_{\epsilon, \lambda, \Gamma_0})\qquad\qquad u_t + \frac{\partial u}{\partial \vec{n}} =g$ on $\Gamma_1 \times (0,T)$

$\qquad\qquad u=0$ on $\Gamma_0 \times (0,T)$

where $0< \epsilon \leq \epsilon_0$, $\Omega \subset \mathbb R^N$ is a regular open connected set, $\lambda \geq 0$ and $\Gamma = \Gamma_0\cup \Gamma_1$ is a partition of the boundary of $\Omega$. We will also consider the case where $\Gamma_0$ is empty (see below for more precise assumptions on $\lambda$, $\Omega$ and $\Gamma_0$, $\Gamma_1$).
For this problem the corresponding formal singular perturbation at $\epsilon =0$ is

$\qquad\qquad \qquad\qquad -\Delta u + \lambda u =f$ on $\Omega \times (0,T)$

$(P_{0, \lambda, \Gamma_0}) \qquad\qquad u_t + \frac{\partial u}{\partial \vec{n}} =g$ on $\Gamma_1 \times (0,T)$

$\qquad\qquad u=0$ on $\Gamma_0 \times (0,T)$

We are here concerned with the well possedness of both problems for the non--homogeneous case, i.e. $f=f(t,x)$, $g=g(t,x)$, and with the convergence, as $\epsilon$ approaches $0$, of the solutions of $(P_{\epsilon, \lambda, \Gamma_0})$ to solutions of $(P_{0, \lambda, \Gamma_0})$.

1995, 1(3): 347-369 doi: 10.3934/dcds.1995.1.347 +[Abstract](966) +[PDF](262.7KB)
Abstract:
We characterize the $\Gamma$-convergence of one-dimensional integral functionals with bulk and jump-part energies, by means of a suitable convergence of the integrands.
1995, 1(3): 371-388 doi: 10.3934/dcds.1995.1.371 +[Abstract](1075) +[PDF](222.1KB)
Abstract:
In this paper, we study an optimal control problem for semilinear evolution equations with a mixed constraint of the state and the control. One of the motivations is the problem with a state dependent control domain. Our main result is the Pontryagin type necessary conditions for the optimal controls. The main tools we use are the Ekeland variational principle and the spike perturbation technique.
1995, 1(3): 389-400 doi: 10.3934/dcds.1995.1.389 +[Abstract](820) +[PDF](190.9KB)
Abstract:
The paper studies periodic solutions of Hamiltonian systems. It states that a range of periods for such solutions can be obtained by diiferentiation of the function of constrained critical values with respect to the variable constraint level. It also shows that when a Hamiltonian is equal to a positive quadratic form plus an oscillatory term, there exist infinitely many solutions with the same period.
1995, 1(3): 401-420 doi: 10.3934/dcds.1995.1.401 +[Abstract](935) +[PDF](235.6KB)
Abstract:
We study the variational convergence, as $\h \rightarrow \infty$, of a sequence of optimal control problems $(\mathcal{P}_h)$ with abstract state equations $A_h(y)=B_h(u)$, where $A_h$ are $G$-converging and the operators $B_h$ acting on the controls are supposed continuously converging, or nonlinear but local, or linear but possibly nonlocal.
1995, 1(3): 421-448 doi: 10.3934/dcds.1995.1.421 +[Abstract](1066) +[PDF](258.8KB)
Abstract:
A fairly general class of nonlinear evolution equations with a self-adjoint or non self-adjoint linear operator is considered, and a family of approximate inertial manifolds (AIMs) is constructed. Two cases are considered: when the spectral gap condition (SGC) is not satisfied and an exact inertial manifold is not known to exist the construction is such that the AIMs have exponential order, while when the SGC is satisfied (and hence there exists an exact inertial manifold) the construction is such that the AIMs converge exponentially to the exact inertial manifold.
1995, 1(3): 449-461 doi: 10.3934/dcds.1995.1.449 +[Abstract](1285) +[PDF](194.3KB)
Abstract:
We give examples of a bounded solution whose gradient blows up in a finite time but it stays bounded on the boundary for a class of quasilinear parabolic equations with zero boundary data. The method reflects a geometric argument for curve evolution equations.

2018  Impact Factor: 1.143