arXiv reader

Regularity to Timoshenko’s System with Thermoelasticity of Type III with Fractional Damping

Filomena Barbosa Rodrigues Mendes, Lesly Daiana Barbosa Sobrado and Fredy Maglorio Sobrado Suárez

arXiv
2211.10816v2 · PDF · 20 August 2023
Categories
math.AP
Journal
Atena Editora, Editora Atena, as a chapter of the book "Ci\^encias exatas e da terra: teorias e princ\'ipios 2", on August 2, (2023), Pages 1--29
DOI
10.22533/at.ed.3742302081
Licence
CC BY 4.0 The site shows the text.

Abstract

The article, presents the study of the regularity of two thermoelastic beam systems defined by the Timoshenko beam model coupled with the heat conduction of Green-Naghdiy theory of type III, both mathematical models are differentiated by their coupling terms that arise as a consequence of the constitutive laws initially considered. The systems presented in this work have 3 fractional dampings: \mu_1(-\Delta)^\tau \phi_t, \mu_2(-\Delta)^\sigma \psi_t and K(-\Delta)^\xi \theta_t, where \phi,\psi and \theta are transverse displacement, rotation angle and empirical temperature of the bean respectively and the parameters (\tau,\sigma,\xi)\in [0,1]^3. It is noted that for values 0 and 1 of the parameter \tau, the so-called frictional or viscous damping will be faced, respectively. The main contribution of this article is to show that the corresponding semigroup S_i(t)=e^{\mathcal{B}_it}, with i=1,2, is of Gevrey class s>\frac{r+1}{2r} for r=\min \{\tau,\sigma,\xi\} for all (\tau,\sigma,\xi )\in R_{CG}:= (0, 1)^3. It is also showed that S_1(t)=e^{\mathcal{B}_1t} is analytic in the region R_{A_1}:=\{(\tau,\sigma, \xi )\in [\frac{1}{2},1]^3\} and S_2(t)=e^{\mathcal{B}_2t} is analytic in the region R_{A_2}:=\{(\tau,\sigma, \xi )\in [\frac{1}{2},1]^3/ \tau=\xi\}.

Items 28

Theorem

Theorem 1

MIJ5

Let \mathcal{B}_i be a linear operator with domain \mathcal{D}(\mathcal{B}_i) dense in a Hilbert space \mathcal{H}. If \mathcal{B}_i is dissipative and 0\in\rho( \mathcal{B}_i), the resolvent set of \mathcal{B}_i, then \mathcal{B}_i is the generator of a C_0- semigroup of contractions on \mathcal{H}.

Theorem

Theorem 2

X9RS

Given U_0\in\mathcal{H} there exists a unique weak solution U to the problem (?) satisfying

U\in C([0, +\infty), \mathcal{H}).

Futhermore, if U_0\in D(\mathcal{B}_i^k), \; k\in\mathbb{N}, then the solution U of (?) satisfies

U\in \bigcap_{j=0}^kC^{k-j}([0,+\infty), D(\mathcal{B}_i^j).
Lemma

Lemma 3

KRN8

Let S(t)=e^{\mathcal{B}_1t} be a C_0-semigroup on contractions on Hilbert space \mathcal{H}=\mathcal{H}_1, the solutions of the system (?)–(?) and (?)–(?) satisfy the inequality

\limsup_{|\lambda|\to\infty}\|(i\lambda I-\mathcal{B}_1) ^{-1}\|_{\mathcal{L}(\mathcal{H})}<\infty.
Proof of Lemma 3

To show the (3) inequality, it suffices to show that, given \delta>0 there exists a constant C_\delta>0 such that the solutions of the system (?)–(?) and (?)–(?) for |\lambda|>\delta satisfy the inequality

\dfrac{\|U\|_{\mathcal{H}}}{\|F\|_{\mathcal{H}}}\leq C_\delta\Longleftrightarrow \|U\|_{\mathcal{H}}=\|(i\lambda I-\mathcal{B}_1)^{-1} F\|_{\mathcal{H}} \leq C_\delta \|F\|_\mathcal{H}.

If \lambda\in \mathbb{R} and F=(f^1,f^2,f^3,f^4, f^5, f^6)\in \mathcal{H} then the solution U=(\phi,\Phi,\psi, \Psi,\theta,\Theta)\in\hbox{D}(\mathcal{B}_1) the resolvent equation (i\lambda I- \mathcal{B}_1)U=F can be written in the form

\begin{aligned} i\lambda \phi-\Phi &= f^1\quad\rm{in}\quad D(A^\frac{1}{2}),\\ i\lambda \Phi-\frac{\kappa}{\rho_1}(\phi_x+\psi)_x+\frac{\mu_1}{\rho_1}A^\tau \Phi&= f^2\quad\rm{in}\quad D(A^0),\\ i\lambda \psi-\Psi &=f^3\quad \rm{in}\quad D(A^\frac{1}{2}),\\ i\lambda \Psi+\dfrac{b}{\rho_2}A\psi+\dfrac{\kappa}{\rho_2}(\phi_x+\psi)+\frac{\beta}{\rho_2}\Theta_x+\dfrac{\mu_2}{\rho_2}A^\sigma \Psi &= f^4 \quad\rm{in}\quad D(A^0),\\ i\lambda\theta-\Theta&=f^5\quad{ \rm in}\quad D(A^\frac{1}{2}),\\ i\lambda\Theta+\dfrac{\delta}{\rho_3}A\theta+\dfrac{\gamma}{\rho_3}\Psi_x+\dfrac{K}{\rho_3}A^\xi\Theta&= f^6\quad {\rm in}\quad D(A^0). \end{aligned}

Using the fact that the operator \mathcal{B}_1 is dissipative, result in

\begin{gathered} \beta\gamma\mu_1\|A^\frac{\tau}{2}\Phi\|^2+\kappa\gamma\mu_2\|A^\frac{\sigma}{2} \Psi\|^2+\dfrac{\beta\kappa K^2}{\rho_3}\|A^\frac{\xi}{2}\Theta\|^2 = \text{Re}\langle{(i\lambda -\mathcal{B}_1)U},{U}\rangle\\= \text{Re}\langle{F},{U}\rangle\leq \|F\|_\mathcal{H}\|U\|_\mathcal{H}. \end{gathered}

On the other hand, performing the duality product of (?) for \beta\gamma\rho_1 \phi, and remembering that the operators A^\nu for all \nu\in\mathbb{R} are self-adjoint, resulting in

\begin{gathered} \beta\kappa\gamma\langle{(\phi_x+\psi)},{\phi_x}\rangle= \beta\gamma\rho_1\|\Phi\|^2+\beta\gamma\rho_1\langle{\Phi},{f^1}\rangle\\ -i\lambda\mu_1\beta\gamma\|A^\frac{\tau}{2}\phi\|^2 +\mu_1\beta\gamma\langle{A^\frac{\tau}{2} f^1},{A^\frac{\tau}{2} \phi}\rangle +\beta\gamma\rho_1\langle{f^2},{\phi}\rangle, \end{gathered}

now performing the duality product of (?) for \beta\gamma\rho_2 \psi, and remembering that the operators A^\nu for all \nu\in\mathbb{R} are self-adjoint, leading to

\begin{gathered} \beta\kappa\gamma\langle{(\phi_x+\psi)},{\psi}\rangle+b\beta\gamma\|A^\frac{1}{2}\psi\|^2=\beta\gamma\rho_2\|\Psi\|^2 +\beta\gamma\rho_2\langle{\Psi},{f^3}\rangle\\ +\beta^2\gamma\langle{\Theta},{\psi_x}\rangle-i\lambda\beta\gamma\mu_2\|A^\frac{\sigma}{2}\psi\|^2+\beta\gamma\mu_2\langle{A^\frac{\sigma}{2}f^3},{A^\frac{\sigma}{2}\psi}\rangle+\beta\gamma\rho_2\langle{f^4},{\psi}\rangle. \end{gathered}

Adding the last 2 equations, result in

\begin{gathered} \beta\kappa\gamma\|\phi_x+\psi\|^2+b\beta\gamma\|A^\frac{1}{2}\psi\|^2= \beta\gamma\rho_1\|\Phi\|^2+\beta\gamma\rho_2\|\Psi\|^2 -i\lambda\beta\gamma\{ \mu_1\|A^\frac{\tau}{2}u\|^2\\ +\mu_2\|A^\frac{\sigma}{2}\psi\|^2\} +\beta\gamma\rho_1\langle{\Phi},{f^1}\rangle+\beta\gamma\rho_2\langle{\Psi},{f^3}\rangle\\ +\beta\gamma\mu_1\langle{A^\frac{\tau}{2}f^1},{A^\frac{\tau}{2}\phi}\rangle +\beta\gamma\mu_2\langle{A^\frac{\sigma}{2}f^3},{A^\frac{\sigma}{2}\phi}\rangle\\ +\beta\gamma\rho_1\langle{f^2},{\phi}\rangle+\beta\gamma\rho_2\langle{f^4},{\psi}\rangle+\beta^2\gamma\langle{\Theta},{\psi_x}\rangle. \end{gathered}

Taking real part, using norm \|F\|_\mathcal{H} and \|U\|_\mathcal{H} and applying Cauchy-Schwarz and Young, for \varepsilon>0 exists C_\varepsilon>0 which does not depend on \lambda, such that

\begin{gathered} \beta\kappa\gamma\|\phi_x+\psi\|^2+b\beta\gamma\|A^\frac{1}{2}\psi\|^2 \leq \beta\gamma\rho_1\|\Phi\|^2+\beta\gamma\rho_2\|\Psi\|^2+\varepsilon\|A^\frac{1}{2}\psi\|^2+C_\varepsilon\|\Theta\|^2 +C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}. \end{gathered}

From estimative (?) and the fact 0\leq\frac{\tau}{2}, \; 0\leq\frac{\sigma}{2} and 0\leq\frac{\xi}{2}, the continuous embedding D(A^{\theta_2}) \hookrightarrow D(A^{\theta_1}),\;\theta_2>\theta_1, lead to

\begin{aligned} \beta\kappa\gamma\|\phi_x+\psi\|^2+b\beta\gamma\|A^\frac{1}{2}\psi\|^2&\leq& C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}. \end{aligned}

On the other hand, performing the duality product of (?) for \beta\kappa\rho_3 \theta, and remembering that the operators A^\nu for all \nu\in\mathbb{R} are self-adjoint and from (?), result in

\begin{gathered} \beta\kappa\delta\|A^\frac{1}{2}\theta\|^2=\delta\beta\kappa\|\Theta\|^2+\beta\kappa\rho_3\langle{\Theta},{f^5}\rangle+\beta\kappa\gamma\langle{\Psi},{\theta_x}\rangle\\ -i\kappa\beta K\lambda\|A^\frac{\xi}{2}\theta\|^2+\kappa\beta K\langle{A^\frac{\xi}{2}f^5},{A^\frac{\xi}{2}\theta}\rangle+\beta\kappa\rho_3\langle{f^6},{\theta}\rangle \end{gathered}

Taking real part and by using the inequalities Cauchy-Schwarz and Young and norms \|F\|_\mathcal{H} and \|U\|_\mathcal{H}, from |\lambda|>1, for \varepsilon>0 exists C_\varepsilon>0 which does not depend on \lambda such that

\begin{gathered} \beta\kappa\delta\|A^\frac{1}{2}\theta\|^2 \leq C\{\|\Theta\|^2+\|\Theta\|\|f^5\|+\|A^\frac{\xi}{2}f^5\|\|A^\frac{\xi}{2}\theta\|+\|f^6\|\|\theta\|\}+C_\varepsilon\|\Psi\|^2+\varepsilon\|A^\frac{1}{2}\theta\|^2. \end{gathered}

Then, from estimative (?) and the fact 0\leq\frac{\sigma}{2} and 0\leq\frac{\xi}{2}\leq\frac{1}{2} the continuous embedding D(A^{\theta_2}) \hookrightarrow D(A^{\theta_1}),\;\theta_2>\theta_1, lead to

\beta\kappa\delta\|A^\frac{1}{2}\theta\|^2\leq C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}.

Finally, from estimates (?), (?) and (?), the proof of this lemma is completed.

Lemma

Lemma 4

UKLP

Let \delta> 0. There exists a constant C_\delta > 0 such that the solutions of the system (?)–(?) and (?)–(?) for |\lambda|\geq \delta satisfy the inequalities

\begin{aligned} \hspace*{-0.5cm}(i)\; |\lambda|\beta\gamma[\kappa\|\phi_x+\psi\|^2+b\|A^\frac{1}{2}\psi\|^2]\hspace*{-0.25cm} &\leq &\hspace*{-0.25cm} |\lambda|\beta\gamma[\rho_1\|\Phi\|^2+\rho_2\|\Psi\|^2] +C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}.\\ \\ (ii)\; \beta\kappa\delta|\lambda|\|A^\frac{1}{2}\theta\|^2 &\leq & \beta\kappa\rho_3|\lambda|\|\Theta\|^2+ C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}. \end{aligned}
Used by Proof, Proof
Proof of Lemma 4

Item (i): Performing the duality product of (?) for \beta\gamma\rho_1\lambda \phi, and remembering that the operators A^\nu for all \nu\in \mathbb{R} are self-adjunct, result in

\begin{aligned} \beta\kappa\gamma\lambda\langle{(\phi_x+\psi)},{\phi_x}\rangle& =&\beta\gamma\rho_1\lambda\|\Phi\|^2+\beta \gamma\rho_1\langle{\lambda \Phi},{f^1}\rangle-i\mu_1\beta\gamma\|A^\frac{\tau}{2}\Phi\|^2\\ & & -i\mu_1\beta\gamma\langle{A^\tau \Phi},{f^1}\rangle +i\beta\gamma\rho_1\langle{f^2},{\Phi}\rangle+i\beta\gamma\rho_1\langle{f^2},{f^1}\rangle. \end{aligned}

Now performing the duality product of (?) for \beta\gamma\rho_2\lambda \psi, and remembering that the operators A^\nu for all \nu\in \mathbb{R} are self-adjunct leads to

\begin{aligned} \lambda\beta\gamma[\kappa\langle{(\phi_x+\psi)},{\psi}\rangle+ b\|A^\frac{1}{2}\psi\|^2] \hspace*{-0.3cm}& = &\hspace*{-0.3cm} \beta\gamma\rho_2\lambda\|\Psi\|^2+\beta\gamma\rho_2\langle{\lambda \Psi},{f^3}\rangle+i\beta^2\gamma\langle{\Theta},{\psi_x}\rangle \\ & &\hspace*{-0.3cm} +i\beta^2\gamma\langle{\Theta},{f_x^3}\rangle - i\mu_2\beta\gamma\|A^\frac{\sigma}{2}\Psi\|^2-i\mu_2\beta\gamma\langle{A^\sigma \Psi},{f^3}\rangle\\ & & \hspace*{-0.3cm} +i\beta\gamma\rho_2\langle{f^4},{\Psi}\rangle +i\beta\gamma\rho_2\langle{f^4},{f^3}\rangle. \end{aligned}

Adding the last 2 equations, result in

\begin{gathered} \lambda\beta\gamma[\kappa\|\phi_x+\psi\|^2+b\|A^\frac{1}{2}\psi\|^2] = \lambda\beta\gamma[\rho_1\|\Phi\|^2+\rho_2\|\Psi\|^2] -i\beta\gamma\{\mu_1\|A^\frac{\tau}{2}\Phi\|^2\\+\mu_2\|A^\frac{\sigma}{2}\Psi\|^2\}+\beta\gamma\rho_1\langle{\lambda \Phi},{f^1}\rangle+\beta\gamma\rho_2\langle{\lambda \Psi},{f^3}\rangle-i\beta\gamma\{\mu_1 \langle{A^\tau \Phi},{f^1}\rangle\\ +\mu_2\langle{A^\sigma \Psi},{f^3}\rangle \} +i\beta\gamma\rho_1\langle{f^2},{\Phi}\rangle +i\beta\gamma\rho_2\langle{f^4},{\Psi}\rangle +i\beta\gamma\rho_1\langle{f^2},{f^1}\rangle\\ +i\beta\gamma\rho_2\langle{f^4},{f^3}\rangle+i\beta^2\gamma\langle{\Theta},{\psi_x}\rangle+i\beta^2\gamma\langle{\Theta},{f_x^3}\rangle. \end{gathered}

On the other hand, from (?) and (?), result in

\begin{gathered} \beta\gamma\{\rho_1\langle{\lambda \Phi},{f^1}\rangle+\rho_2\langle{\lambda \Psi},{f^3}\rangle\} =i\beta\gamma\{\kappa\langle{(\phi_x+\psi)},{f^1_x}\rangle\\ +\mu_1\langle{A^\tau \Phi},{f^1}\rangle-\rho_1\langle{f^2},{f^1}\rangle +b\langle{A^\frac{1}{2}\psi},{A^\frac{1}{2}f^3}\rangle+\kappa\langle{(\phi_x+\psi)},{f^3}\rangle\\ -\beta\langle{\Theta},{f_x^3}\rangle+\mu_2\langle{A^\sigma \Psi},{f^3}\rangle-\rho_2\langle{f^4},{f^3}\rangle\}. \end{gathered}

Using the identity (?) in the (?) equation and simplifying, lead to

\begin{gathered} \lambda\beta\gamma[\kappa\|\phi_x+\psi\|^2+b\|A^\frac{1}{2}\psi\|^2] = \lambda\beta\gamma[\rho_1\|\Phi\|^2+\rho_2\|\Psi\|^2] \\-i\{\mu_1\|A^\frac{\tau}{2}\Phi\|^2 +\mu_2\|A^\frac{\sigma}{2}\Psi\|^2\} +i\beta\gamma\kappa\langle{\phi_x},{f_x^1}\rangle+i\beta\gamma\kappa\langle{\psi},{f_x^1}\rangle\\+ib\beta\gamma\langle{A^\frac{1}{2}\psi},{A^\frac{1}{2}f^3}\rangle i\beta\gamma\kappa\langle{\phi_x},{f^3}\rangle+i\beta\gamma\kappa\langle{\psi},{f^3}\rangle\\ +i\beta\gamma\rho_1\langle{f^2},{\Phi}\rangle +i\beta\gamma\rho_2\langle{f^4},{\Psi}\rangle+i\beta^2\gamma\langle{\Theta},{\psi_x}\rangle. \end{gathered}

As for \varepsilon>0 exists C_\varepsilon> such that |i\beta^2\gamma\langle{\Theta},{\psi_x}\rangle|\leq \varepsilon\|A^\frac{1}{2}\psi\|^2+C_\varepsilon\|\Theta\|^2, taking the real part, and applying the Cauchy-Schwarz and Young inequalities, estimative (Proof) of Lemma(3) and using the definitions of the F and U norm in (?) the proof of item (i) this lemma is completed.

On the other hand, performing the duality product of (?) for \beta\kappa\rho_3\lambda \theta, and remembering that the operators A^\theta for all \theta\in \mathbb{R} are self-adjunct, result in

\begin{gathered} \beta\kappa\delta\lambda\|A^\frac{1}{2}\theta\|^2=\beta\kappa\rho_3\langle{\lambda \Theta},{i\lambda\theta}\rangle-\beta\kappa\gamma\langle{\psi_x},{\lambda\theta}\rangle +\beta\kappa K\langle{A^\xi\Theta},{\lambda\theta}\rangle+\beta\kappa\rho_3\langle{f^6},{\lambda\theta}\rangle. \end{gathered}

As

\begin{aligned} \rho_3\langle{\lambda \Theta},{i\lambda\theta}\rangle&= \rho_3\lambda\|\Theta\|^2+i\delta\langle{A^\frac{1}{2}\theta},{A^\frac{1}{2}f^5}\rangle-i\gamma\langle{\Psi},{f^5_x}\rangle +i K\langle{A^\xi\Theta},{f^5}\rangle -i\rho_3\langle{f^6},{f^5}\rangle.\\ -\gamma\langle{\psi_x},{\lambda\theta}\rangle &=-i\gamma\langle{\psi_x},{\Theta}\rangle-i\gamma\langle{\psi_x},{f^5}\rangle.\\ K\langle{A^\xi\Theta},{\lambda\theta}\rangle &= -i K\|A^\frac{\xi}{2}\Theta\|^2-i K\langle{A^\xi\Theta},{f^5}\rangle. \\ \rho_3\langle{f^6},{\lambda\theta}\rangle &= & i\rho_3\langle{f^6},{\Theta}\rangle+i\rho_3\langle{f^6},{f^5}\rangle. \end{aligned}

Using (?)–(?) in (?), lead to

\begin{gathered} \beta\kappa\delta\lambda\|A^\frac{1}{2}\theta\|^2 = \beta\kappa\rho_3\lambda\|\Theta\|^2+i\beta\kappa\delta\langle{A^\frac{1}{2}\theta},{A^\frac{1}{2}f^5}\rangle-i\beta\kappa\gamma\langle{\Psi},{f^5_x}\rangle-i\beta\kappa\gamma\langle{\psi_x},{\Theta}\rangle\\ -i\beta\kappa\gamma\langle{\psi_x},{f^5}\rangle-i\beta\kappa K\|A^\frac{\xi}{2}\Theta\|^2 +i\beta\kappa\rho_3\langle{f^6},{\Theta}\rangle. \end{gathered}

Taking real part of the equation (?) and using the inequalities Cauchy-Schwarz and Young and norms \|F\|_\mathcal{H} and \|U\|_\mathcal{H}, from |\lambda|>1, for \varepsilon>0 exists C_\varepsilon>0 such that

\begin{gathered} \beta\kappa\delta|\lambda|\|A^\frac{1}{2}\theta\|^2 \leq\beta\kappa\rho_3|\lambda|\|\Theta\|^2+ C\{\|A^\frac{1}{2}\theta\|\|A^\frac{1}{2}f^5\|+\|\Psi\|\|f^5_x\|\\ +\|\psi_x\|^2+\|\Theta\|^2+\|\psi_x\|\|f^5\|+\|f^6\|\|\theta\| \}. \end{gathered}

Then, from Lemma(3) and as \|\psi_x\|^2=\|A^\frac{1}{2}\psi\|^2, the proof of item (ii) of this lemma is finished.

Definition

Definition 5

1M0I

Let t_0\geq 0 be a real number. A strongly continuous semigroup S(t), defined on a Banach space \mathcal{H}, is of Gevrey class s > 1 for t > t_0, if S(t) is infinitely differentiable for t > t_0, and for every compact set K \subset (t_0,\infty) and each \mu > 0, there exists a constant C = C(\mu, K) > 0 such that

||S^{(n)}(t)||_{\mathcal{L}( \mathcal{H})} \leq C\mu ^n(n!)^s, \text{ for all } \quad t \in K, n = 0,1,2...
Theorem

Theorem 6

REZ8

Let S(t) be a strongly continuous and bounded semigroup on a Hilbert space \mathcal{H}. Suppose that the infinitesimal generator \mathcal{B} of the semigroup S(t) satisfies the following estimate, for some 0 < \eta < 1:

\lim\limits_{|\lambda|\to\infty} \sup |\lambda |^\eta ||(i\lambda I-\mathcal{B})^{-1}||_{\mathcal{L}( \mathcal{H})} < \infty.

Then S(t) is of Gevrey class s for t>0, for every s>\dfrac{1}{\eta}.

Used by Proof, Proof
Theorem

Theorem 7

LOPJ

The semigroup S(t)=e^{\mathcal{B}_1t} associated to the system (?)–(?) and (?)–(?) is of Gevrey class s>\frac{r+1}{2r} for r=\min \{\tau,\sigma,\xi\} for all (\tau,\sigma,\xi )\in R_{CG}:= (0, 1)^3.

Proof of Theorem 7

From the resolvent equation F=(i\lambda I-\mathcal{B}_1)U for \lambda\in\mathbb{R}, U=(i\lambda I-\mathcal{B}_1)^{-1}F. Furthermore, show (6) of theorem(6) it is enough to show, that for \varepsilon>0 exists C_\delta>0, such that:

\dfrac{|\lambda|^\frac{2r}{r+1} \|(i\lambda-\mathcal{B}_1)^{-1}F\|_\mathcal{H}}{\|F\|_\mathcal{H}}\leq C_\delta \Longleftrightarrow |\lambda |^\frac{2r}{r+1} \|U\|_\mathcal{H}\leq C_\delta \|F\|_\mathcal{H}, \qquad F\in\mathcal{H}\quad{\rm and}\quad C_\delta>0.

Equivalent to

|\lambda |^\frac{2r}{r+1}\|U\|_\mathcal{H}^2\leq C_\delta\{\|F\|_\mathcal{H}\|U\|_\mathcal{H}\}\quad\rm{for}\quad \frac{2r}{r+1} \in(0,1).

where r=\min\{\tau,\sigma,\xi\}, for all (\tau,\sigma,\xi)\in (0,1)^3.

Next, [|\lambda|^\frac{2\tau}{1+\tau}\|\Phi\|+|\lambda|^\frac{2\sigma}{1+\sigma}\|\Psi\|+|\lambda|^\frac{2\xi}{1+\xi}\|\Theta\|] will be estimated.

Let’s start by estimating the term |\lambda|^\frac{2\tau}{1+\tau}\|\Phi\|: It is assume that |\lambda|>1, some ideas could be borrowed from [LiuR95]. Set \Phi=\Phi_1+\Phi_2, where \Phi_1\in D(A) and \Phi_2\in D(A^0), with

i\lambda \Phi_1+A \Phi_1=f^2, \hspace{3cm} i\lambda \Phi_2=\dfrac{\kappa}{\rho_1}A\phi+\dfrac{\kappa}{\rho_1}\psi_x-\dfrac{\mu_1}{\rho_1}A^\tau \Phi+A\Phi_1.

Firstly, applying in the product duality the first equation in (?) by \Phi_1, then by A\Phi_1 and recalling that the operator A is self-adjoint, resulting in

|\lambda|\|\Phi_1\|+|\lambda|^\frac{1}{2}\|A^\frac{1}{2}\Phi_1\|+\|A\Phi_1\|\leq C\|F\|_\mathcal{H}.

Applying the A^{-\frac{1}{2}} operator on the second equation of (?), result in

i\lambda A^{-\frac{1}{2}}\Phi_2= \dfrac{\kappa}{\rho_1}A^\frac{1}{2}\phi+\dfrac{\kappa}{\rho_1}A^{-\frac{1}{2}}\psi_x-\dfrac{\mu_1}{\rho_1}A^{\tau-\frac{1}{2}} \Phi+A^\frac{1}{2}\Phi_1,

then, as \|A^{-\frac{1}{2}}\psi_x\|^2=\langle{-A^{-\frac{1}{2}}\psi_{xx}},{A^{-\frac{1}{2}}\psi}\rangle=\langle{A^\frac{1}{2}\psi},{A^{-\frac{1}{2}}\psi}\rangle=\|\psi\|^2 and \tau-\frac{1}{2}\leq \frac{\tau}{2} taking into account the continuous embedding D(A^{\theta_2}) \hookrightarrow D(A^{\theta_1}),\;\theta_2>\theta_1, result in

|\lambda|\|A^{-\frac{1}{2}} \Phi_2\|\leq C\{\|A^\frac{1}{2}\phi\|+ \|\psi\|+\|A^\frac{\tau}{2}\Phi\|\}+\|A^\frac{1}{2}\Phi_1\|,

using (Proof), (?) and estimative (?) (\|A^\frac{1}{2}\Phi\|^2\leq C|\lambda|^{-1}\|F\|^2_\mathcal{H}), lead to

|\lambda|^2\|A^{-\frac{1}{2}}\Phi_2\|^2\leq C\{\|F\|_\mathcal{H}\|U\|_\mathcal{H} +|\lambda|^{-1}\|F\|^2_\mathcal{H} \}.

Equivalently

\|A^{-\frac{1}{2}}\Phi_2\|^2\leq C|\lambda|^{-\frac{4\tau+2}{\tau+1}}\{ |\lambda|^\frac{2\tau}{\tau+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}\quad {\rm for}\quad 0\leq\tau\leq 1.

On the other hand, from \Phi_2=\Phi-\Phi_1, (?) and as \frac{\tau}{2}\leq\frac{1}{2} the inequality of (?), result in

\begin{aligned} \|A^\frac{\tau}{2} \Phi_2\|^2 & \leq & C\{ \|A^\frac{\tau}{2} \Phi\|^2+\|A^\frac{\tau}{2}\Phi_1\|^2 \}\\ &\leq & C\{\|F\|_\mathcal{H}\|U\|_\mathcal{H}+|\lambda|^{-1}\|F\|^2_\mathcal{H} \}\leq C|\lambda|^{-\frac{2\tau}{\tau+1}}\{ |\lambda|^\frac{2\tau}{\tau+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}. \end{aligned}

By Lions’ interpolations inequality 0\in [-\frac{1}{2},\frac{\tau}{2}], result in

\|\Phi_2\|^2\leq C(\|A^{-\frac{1}{2}}\Phi_2\|^2)^\frac{\tau}{1+\tau}(\|A^\frac{\tau}{2}\Phi_2\|^2)^\frac{1}{1+\tau}.

Then, using (?) and (?) in (?), for |\lambda|>1, result in

\|\Phi_2\|^2\leq C|\lambda|^{-\frac{4\tau}{1+\tau}}\{|\lambda|^\frac{2\tau}{\tau+1} \|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}.

Therefore, as \|\Phi\|^2\leq \|\Phi_1\|^2+ \|\Phi_2\|^2 from first inequality of (?), (?), we get

\|\Phi\|^2\leq C|\lambda|^{-2}\|F\|^2_\mathcal{H}+ C|\lambda|^{-\frac{4\tau}{1+\tau}}\{|\lambda|^\frac{2\tau}{\tau+1} \|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\},

and as for 0\leq\tau\leq 1 we have |\lambda|^{-2}\leq |\lambda|^\frac{-4\tau}{1+\tau}, result in

|\lambda|\|\Phi\|^2\leq C_\delta|\lambda|^\frac{1-3\tau}{1+\tau}\{|\lambda|^\frac{2\tau}{1+\tau}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}\qquad\rm{for} \qquad 0\leq\tau\leq 1.

On the other hand, let’s now estimate the missing term |\lambda|^\frac{2\sigma}{1+\sigma}\|\Psi\|: It is assumed that |\lambda|>1. Set \Psi=\Psi_1+\Psi_2, where \Psi_1\in D(A) and \Psi_2\in D(A^0), with

\begin{gathered} i\lambda \Psi_1+A \Psi_1=f^4 \qquad {\rm and}\qquad i\lambda \Psi_2=-\dfrac{b}{\rho_2}A\psi -\dfrac{\kappa}{\rho_2}\phi_x-\dfrac{\kappa}{\rho_2}\psi-\dfrac{\beta}{\rho_2}\Theta_x-\dfrac{\mu_2}{\rho_2}A^\sigma \Psi+A\Psi_1. \end{gathered}

Firstly, applying in the product duality the first equation in (?) by \Psi_1, then by A\Psi_1 and recalling that the operator A is self-adjoint, resulting in

|\lambda|\|\Psi_1\|+|\lambda|^\frac{1}{2}\|A^\frac{1}{2}\Psi_1\|+\|A\Psi_1\|\leq C\|F\|_\mathcal{H}.

As follows from the second equation in (?) that

i\lambda A^{-\frac{1}{2}}\Psi_2= -\dfrac{b}{\rho_2}A^\frac{1}{2}\psi-\dfrac{\kappa}{\rho_2}A^{-\frac{1}{2}}\phi_x-\dfrac{\kappa}{\rho_2}A^{-\frac{1}{2}}\psi-\dfrac{\beta}{\rho_2}A^{-\frac{1}{2}}\Theta_x-\dfrac{\mu_2}{\rho_2}A^{\sigma-\frac{1}{2}}\Psi+A^\frac{1}{2}\Psi_1,

then, as \|A^{-\frac{1}{2}}\phi_x\|^2=\langle{-A^{-\frac{1}{2}}\phi_{xx}},{A^{-\frac{1}{2}}\phi}\rangle=\|\phi\|^2, \|A^{-\frac{1}{2}}\Theta_x\|^2=\|\Theta\|^2 and \sigma-\frac{1}{2}\leq \frac{\sigma}{2} taking into account the continuous embedding D(A^{\theta_2}) \hookrightarrow D(A^{\theta_1}),\;\theta_2>\theta_1, lead to

|\lambda|\|A^{-\frac{1}{2}} \Psi_2\|\leq C\{ \|A^\frac{1}{2}\phi\|+\|\psi\|+\|\Theta\|+\|A^\frac{\sigma}{2}\Psi\|\}+\|A^\frac{1}{2}\Psi_1\|

Using estimative (Proof) and estimative (?), for |\lambda|>1, result in

\|A^{-\frac{1}{2}}\Psi_2\|^2 \leq C|\lambda|^{-3}\{|\lambda|\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}.

Equivalently

\|A^{-\frac{1}{2}}\Psi_2\|^2 \leq C|\lambda|^{-\frac{4\sigma+2}{\sigma+1}}\{ |\lambda|^\frac{2\sigma}{\sigma+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}\quad {\rm for}\quad 0\leq \sigma\leq 1.

On the other hand, from \Psi_2=\Psi-\Psi_1, (?) and as \frac{\sigma}{2}\leq\frac{1}{2} the inequality (?), result in

\begin{aligned} \|A^\frac{\sigma}{2} \Psi_2\|^ 2& \leq & C\{ \|A^\frac{\sigma}{2} \Psi\|^2+\|A^\frac{\sigma}{2}\Psi_1\|^2\}\\ & \leq & C\{\|F\|_\mathcal{H}\|U\|_\mathcal{H}+|\lambda|^{-1}\|F\|^2_\mathcal{H}\} \leq |\lambda|^{-\frac{2\sigma}{\sigma+1}}\{|\lambda|^\frac{2\sigma}{\sigma+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}. \end{aligned}

Now, by Lions’ interpolations inequality 0\in [-\frac{1}{2},\frac{\sigma}{2}], lead to

\|\Psi_2\|^2 \leq C (\|A^{-\frac{1}{2}}\Psi_2\|^2)^\frac{\sigma}{1+\sigma}(\|A^\frac{\sigma}{2}\Psi_2\|^2)^\frac{1}{1+\sigma}.

Then, using (?) and (?) in (?), for |\lambda|>1, result in

\|\Psi_2\|^2 \leq C|\lambda|^{\frac{-4\sigma}{(1+\sigma)}}\{|\lambda|^\frac{2\sigma}{\sigma+1} \|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|_\mathcal{H}\}.

Therefore, as \|\Psi\|^2\leq \|\Psi_1\|^2+ \|\Psi_2\|^2 from first inequality of (?), (?) and |\lambda|^{-2}\leq|\lambda|^\frac{-4\sigma}{1+\sigma}, result in

|\lambda|\|\Psi\|^2\leq |\lambda|^\frac{1-3\sigma}{1+\sigma}\{|\lambda|^\frac{2\sigma}{\sigma+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}\qquad\rm{for}\qquad 0\leq\sigma\leq 1.

Finally, let’s now estimate the missing term |\lambda|^\frac{2\xi}{1+\xi}\|\Theta\|: Now we assume |\lambda|>1. Set \Theta=\Theta_1+\Theta_2, where \Theta_1\in D(A) and \Theta_2\in D(A^0), with

i\lambda \Theta_1+A \Theta_1=f^6 \qquad{\rm and}\qquad i\lambda \Theta_2=-\dfrac{\delta}{\rho_3}A\theta -\dfrac{\gamma}{\rho_3}\Psi_x-\dfrac{K}{\rho_3}A^\xi \Theta+A\Theta_1.

Firstly, applying in the product duality the first equation in (?) by \Theta_1, then by A\Theta_1 and recalling that the operator A is self-adjoint, resulting in

|\lambda|\|\Theta_1\|+|\lambda|^\frac{1}{2}\|A^\frac{1}{2}\Theta_1\|+\|A\Theta_1\|\leq C\|F\|_\mathcal{H}.

As follows from the second equation in (?) that

i\lambda A^{-\frac{1}{2}}\Theta_2= -\dfrac{\delta}{\rho_3}A^\frac{1}{2}\theta -\dfrac{\gamma}{\rho_3}A^{-\frac{1}{2}}\Psi_x-\dfrac{K}{\rho_3}A^{\xi-\frac{1}{2}} \Theta+A^\frac{1}{2}\Theta_1,

then, as \|A^{-\frac{1}{2}}\Psi_x\|^2=\|\Psi\|^2 and -\frac{1}{2}\leq 0, \xi-\frac{1}{2}\leq \frac{\xi}{2} taking into account the continuous embedding D(A^{\theta_2}) \hookrightarrow D(A^{\theta_1}),\;\theta_2>\theta_1, result in

|\lambda|\|A^{-\frac{1}{2}} \Theta_2\|\leq C\{ \|A^\frac{1}{2}\theta\|+\|A^\frac{\xi}{2}\Theta\|+\|\Psi\|\}+\|A^\frac{1}{2}\Theta_1\|

Using estimates (Proof), (?) and estimative (?), for |\lambda|>1, lead to

\|A^{-\frac{1}{2}}\Theta_2\|^2\leq C|\lambda|^{-\frac{4\xi+2}{\xi+1}}\{|\lambda|^\frac{2\xi}{\xi+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}.

On the other hand, from \Theta_2=\Theta-\Theta_1, (?) and as \frac{\xi}{2}\leq\frac{1}{2} the inequality (?), result in

\begin{aligned} \|A^\frac{\xi}{2} \Theta_2\|^2 & \leq & C\{ \|A^\frac{\xi}{2} \Theta\|^2+\|A^\frac{\xi}{2}\Theta_1\|^2\}\\ & \leq & C|\lambda|^{-\frac{2\xi}{\xi+1}}\{|\lambda|^\frac{2\xi}{\xi+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}. \end{aligned}

Now, by Lions’ interpolations inequality 0\in [-\frac{1}{2},\frac{\xi}{2}], result in

\|\Theta_2\|^2\leq C (\|A^{-\frac{1}{2}}\Theta_2\|^2)^\frac{\xi}{1+\xi}(\|A^\frac{\xi}{2}\Theta_2\|^2)^\frac{1}{1+\xi}.

Then, using (?) and (?) in (?), for |\lambda|>1, result in

\|\Theta_2\|^2\leq C|\lambda|^{\frac{-4\xi}{(1+\xi)}}\{|\lambda|^\frac{2\xi}{\xi+1} \|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|_\mathcal{H}\}.

Therefore, as \|\Theta\|^2\leq \|\Theta_1\|^2+ \|\Theta_2\|^2 from first inequality of (?), estimative (?) and |\lambda|^{-2}\leq |\lambda|^\frac{-4\xi}{1+\xi}, result in

|\lambda|\|\Theta\|^2\leq C_\delta |\lambda|^\frac{1-3\xi}{1+\xi}\{|\lambda|^\frac{2\xi}{\xi+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}\qquad \rm{for}\qquad 0\leq\xi\leq 1.

Using the estimates (?) and (?) in the inequality item (i) of Lemma 4, result in

\begin{gathered} \beta\gamma\kappa\|\phi_x+\psi\|^2+\beta\gamma b\|A^\frac{1}{2}\psi\|^2 \leq C_\delta\big\{ |\lambda|^\frac{-2\tau}{1+\tau}+|\lambda|^\frac{-2\sigma}{1+\sigma}+|\lambda|^{-1}\big\}\|F\|_\mathcal{H}\|U\|_\mathcal{H}\\ +\big\{|\lambda|^{-\frac{4\tau}{\tau+1}}+|\lambda|^{-\frac{4\sigma}{\sigma+1}}\big\}\|F\|^2_\mathcal{H}. \end{gathered}

As \tau\leq \sigma \Longrightarrow -\frac{\tau}{\tau+1}\geq -\frac{\sigma}{\sigma+1} and \tau\geq \sigma \Longrightarrow -\frac{\tau}{\tau+1}\leq -\frac{\sigma}{\sigma+1}, we obtain

\begin{gathered} |\lambda|\{ \beta\gamma\kappa\|\phi_x+\psi\|^2+\beta\gamma b\|A^\frac{1}{2}\psi\|^2 \}\\ \leq C_\delta\left\{ \begin{array}{ccc} |\lambda|^\frac{1-3\tau}{\tau+1}\{|\lambda|^\frac{2\tau}{\tau+1} \|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} & \text{for} & \tau\leq \sigma\\\\ |\lambda|^\frac{1-3\sigma}{\sigma+1}\{ |\lambda|^\frac{2\sigma}{\sigma+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H} \} &\text{for} & \sigma\leq \tau \end{array} \right. \end{gathered}

Furthermore, using estimative (?) in estimate (4) (item (ii) of the Lemma 4) and as \frac{1-3\xi}{\xi+1}+\frac{2\xi}{\xi+1}=\frac{1-\xi}{\xi+1}>0, we have

\begin{gathered} \beta\kappa\delta |\lambda| \|A^\frac{1}{2}\theta\|^2 \leq \beta\kappa\rho_3|\lambda| \|\Theta\|^2+C_\delta|\|F\|_\mathcal{H}\|U\|_\mathcal{H} \leq C_\delta|\lambda|^\frac{1-3\xi}{1+\xi}\{|\lambda|^\frac{2\xi}{1+\xi}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}. \end{gathered}

Finally summing the estimates (?),(?), (?),(?) and (?), we have

|\lambda|\|U\|^2_\mathcal{H}\leq C_\delta\left\{ \begin{array}{ccc} |\lambda|^\frac{1-3\tau}{1+\tau}\{ |\lambda|^\frac{2\tau}{\tau+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} &\text{for} & \tau\leq \sigma\quad{\rm and}\quad \tau\leq \xi, \\\\ |\lambda|^\frac{1-3\sigma}{1+\sigma}\{ |\lambda|^\frac{2\sigma}{\sigma+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} &\text{for} & \sigma \leq \tau\quad{\rm and}\quad \sigma\leq \xi,\\\\ |\lambda|^\frac{1-3\xi}{1+\xi}\{ |\lambda|^\frac{2\xi}{\xi+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} &\text{for} & \xi\leq \sigma\quad{\rm and}\quad \xi\leq \tau. \end{array}\right.

for (\tau,\sigma,\xi)\in [0,1]^3, equivalently

\|U\|^2_\mathcal{H}\leq C_\delta\left\{ \begin{array}{ccc} |\lambda|^\frac{-4\tau}{1+\tau}\{ |\lambda|^\frac{2\tau}{\tau+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} &\text{for} & \tau\leq \sigma\quad{\rm and}\quad \tau\leq \xi, \\\\ |\lambda|^\frac{-4\sigma}{1+\sigma}\{ |\lambda|^\frac{2\sigma}{\sigma+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} &\text{for} & \sigma \leq \tau\quad{\rm and}\quad \sigma\leq \xi,\\\\ |\lambda|^\frac{-4\xi}{1+\xi}\{ |\lambda|^\frac{2\xi}{\xi+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} &\text{for} & \xi\leq \sigma\quad{\rm and}\quad \xi\leq \tau. \end{array}\right.

Finally, applying Young inequality, we have

\left\{ \begin{array}{ccc} |\lambda|^\frac{2\tau}{\tau+1}\|U\|_\mathcal{H}\leq C_\delta|\|F\|_\mathcal{H} &\text{for} & \tau\leq \sigma\quad{\rm and}\quad \tau\leq \xi, \\\\ |\lambda|^\frac{2\sigma}{1+\sigma}\|U\|_\mathcal{H}\leq C_\delta\|F\|_\mathcal{H} &\text{for} & \sigma \leq \tau\quad{\rm and}\quad \sigma\leq \xi,\\\\ |\lambda|^\frac{2\xi}{1+\xi}\|U\|_\mathcal{H}\leq C_\delta \|F\|_\mathcal{H} & \text{for} & \xi\leq \sigma\quad{\rm and}\quad \xi\leq \tau. \end{array}\right.

Therefore, proof of this theorem is finished.

Theorem

Theorem 8

O7GT

Let S(t)=e^{\mathcal{B}t} be C_0-semigroup of contractions on a Hilbert space \mathcal{H}. Suppose that

\rho(\mathcal{B})\supseteq\{ i\lambda/ \lambda\in {\mathbb R} \} \equiv i{\mathbb R}

Then S(t) is analytic if and only if

\limsup\limits_{|\lambda|\to \infty} \|\lambda(i\lambda I-\mathcal{B})^{-1}\|_{\mathcal{L}( \mathcal{H})}<\infty

holds.

Used by Proof, Proof
Lemma

Lemma 9

OA26

The C_0-semigroup of contractions S(t)=e^{\mathcal{B}_1t} satisfies the condition

\rho(\mathcal{B}_1)\supseteq\{ i\lambda/ \lambda\in {\mathbb R} \} \equiv i{\mathbb R}.
Used by Proof, Remark 17
Proof of Lemma 9

Let’s prove i{\mathbb R}\subset\rho( \mathcal{B}_1) by contradiction, it is supposed that i{\mathbb R}\not\subset \rho( \mathcal{B}_1). As 0\in\rho( \mathcal{B}_1) and \rho( \mathcal{B}_1) is open, the highest positive number \lambda_0 is considered such that the interval (-i\lambda_0,i\lambda_0)\subset\rho( \mathcal{B}_1), then i\lambda_0 or -i\lambda_0 is an element of the spectrum \sigma( \mathcal{B}_1). It is supposed that i\lambda_0\in \sigma( \mathcal{B}_1) (if -i\lambda_0\in \sigma( \mathcal{B}_1) the proceeding is similar). Then, for 0<\nu<\lambda_0 there exist a sequence of real numbers (\lambda_n), with 0<\nu\leq\lambda_n<\lambda_0, \lambda_n\to \lambda_0, and a vector sequence U_n=(\phi_n,\Phi_n,\psi_n,\Psi_n, \theta_n,\Theta_n)\in \mathcal{D}( \mathcal{B}_1) with unitary norms, such that

\begin{aligned} \|(i\lambda_n- \mathcal{B}_1) U_n\|_\mathcal{H}=\|F_n\|_\mathcal{H}\to 0, \end{aligned}

as n\to \infty. From (?) for 0\leq\tau\leq 1 and 0\leq\sigma\leq 1, result in

\begin{aligned} \beta\kappa\gamma\|\phi_{xn}+\psi_n\|^2+b\beta\gamma\|A^\frac{1}{2}\psi_n\|^2&\leq& C_\delta\|F_n\|_\mathcal{H}\|U_n\|_\mathcal{H}. \end{aligned}

From (?), for 0\leq\xi\leq 1, result in

\beta\kappa\delta\|A^\frac{1}{2}\theta_n\|^2\leq C_\delta\|F_n\|_\mathcal{H}\|U_n\|_\mathcal{H}.

In addition to the estimative (?), for 0\leq\tau\leq 1 and 0\leq\sigma\leq 1, result in

\begin{aligned} \beta\gamma\mu_1\|A^\frac{\tau}{2}\Phi_n\|^2+\kappa\gamma\mu_2\|A^\frac{\sigma}{2}\Psi_n\|^2+\dfrac{\beta\kappa K}{\rho_3}\|A^\frac{\xi}{2}\Theta_n\|^2\leq C_\delta \|F_n\|_\mathcal{H}\|U_n\|_\mathcal{H}. \end{aligned}

Consequently, \|U_n\|_\mathcal{H}^2 \to 0. Therefore, lead to \|U_n\|_\mathcal{H}\to 0 but this is absurd, since \|U_n\|_\mathcal{H}=1 for all n\in{\mathbb N}. Thus, i{\mathbb R}\subset \rho(\mathcal{B}_1). This completes the proof.

Uses Proof
Lemma

Lemma 10

SKZV

Let \delta> 0. There exists a constant C_\delta > 0 such that the solutions of (?)–(?) and (?)–(?) for |\lambda|\geq \delta satisfy the inequality

\begin{aligned} (i)\quad |\lambda|\|\Phi\|^2\leq C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}\qquad{\rm for}\qquad \frac{1}{2}\leq\tau\leq 1.\\ (ii)\quad |\lambda|\|\Psi\|^2\leq C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}\qquad{\rm for}\qquad \frac{1}{2}\leq\sigma\leq 1.\\ (iii)\quad |\lambda|\|\Theta\|^2\leq C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}\qquad{\rm for}\qquad \frac{1}{2}\leq\xi\leq 1. \end{aligned}
Used by Proof
Proof of Lemma 10

Item (i): of this lemma will be initially shown, performing the duality product of (?) for \dfrac{\rho_1}{\mu_1}A^{-\tau}\lambda \Phi, and recalling that the operator A^\nu is self-adjoint for all \nu\in\mathbb{R}, results in

\begin{aligned} \lambda\|\Phi\|^2\hspace*{-0.3cm} &= \dfrac{\kappa}{\mu_1}\langle{\lambda(-A\phi+\psi_x)},{A^{-\tau} \Phi}\rangle+\dfrac{\rho_1}{\mu_1}\langle{f^2},{\lambda A^{-\tau}\Phi}\rangle-i\dfrac{\rho_1}{\mu_1}\lambda^2\|A^{-\frac{\tau}{2}}\Phi\|^2\\ &=\hspace*{-0.3cm} i\dfrac{\kappa}{\mu_1}\|A^\frac{1-\tau}{2} \Phi\|^2+i\dfrac{\kappa}{\mu_1}\langle{A^\frac{1}{2}f^1},{A^{\frac{1}{2}-\tau}\Phi}\rangle+i\dfrac{\kappa}{\mu_1}\langle{\Psi},{A^{-\tau}\Phi_x}\rangle-i\dfrac{\kappa}{\mu_1}\langle{f_x^3},{A^{-\tau}\Phi}\rangle\\ & &-i\langle{f^2},{\Phi}\rangle -i\dfrac{\kappa}{\mu_1}\langle{f^2},{A^{1-\tau}\phi}\rangle +i\dfrac{\kappa}{\mu_1}\langle{A^{-\tau}f^2},{\psi_x}\rangle +i\dfrac{\rho_1}{\mu_1}\|A^{-\frac{\tau}{2}}f^2\|^2-i\dfrac{\rho_1}{\mu_1}\lambda^2\|A^{-\frac{\theta}{2}}\Phi\|^2. \end{aligned}

Noting that: \|A^{-\tau} \Phi_x\|^2=\langle{A^{-\tau}\Phi_x},{A^{-\tau}\Phi_x}\rangle=\langle{-A^{-\tau}\Phi_{xx}},{A^{-\tau}\Phi}\rangle=\|A^\frac{1-2\tau}{2} \Phi\|^2, taking real part and considering that \frac{1}{2}\leq\tau\leq 1 using estimative (?) and using Cauchy-Schwarz and Young inequalities, for \varepsilon>0 exists C_\varepsilon, such that

\begin{aligned} |\lambda|\|\Phi\|^2 & \leq & C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\varepsilon\|A^{-\tau}\Phi_x\|^2+C_\varepsilon\|\Psi\|^2. \end{aligned}

As 0\leq\frac{\sigma}{2}, then from estimative (?) \|\Psi\|^2\leq C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}. From the continuous embedding for |\lambda|\geq 1, the proof of item (i) of this lemma is finished.

Item (ii): Again similarly, performing the duality product of (?) for \dfrac{\rho_2}{\mu_2}A^{-\sigma}\lambda \Psi, using (?), and recalling the self-adjointness of A^\nu, \nu \in\mathbb{R}, leads to

\begin{aligned} \lambda\|\Psi\|^2 &= i\dfrac{b}{\mu_2}\|A^\frac{1-\sigma}{2}\Psi\|^2+i\dfrac{b}{\mu_2}\langle{A^\frac{1}{2}f^3},{A^{\frac{1}{2}-\sigma}\Psi}\rangle+i\dfrac{\kappa}{\mu_2}\langle{A^{-\sigma}\Phi_x},{\Psi}\rangle+i\dfrac{\kappa}{\mu_2}\langle{f^1_x},{A^{-\sigma} \Psi}\rangle\\ & & +i\dfrac{\kappa}{\mu_2}\|A^{-\frac{\sigma}{2}}\Psi\|^2+i\dfrac{\kappa}{\mu_2}\langle{f^3},{A^{-\sigma}\Psi}\rangle-i\dfrac{b}{\rho_2}\langle{f^4},{A^{1-\sigma}\psi}\rangle-i\dfrac{\kappa}{\mu_2}\langle{f^4},{A^{-\sigma} \phi_x}\rangle\\ & & -i\dfrac{\kappa}{\mu_2}\langle{f^4},{A^{-\sigma}\psi}\rangle-i\dfrac{\beta}{\mu_2}\langle{f^4},{A^{-\sigma}\Theta}\rangle-i\langle{f^4},{\Psi}\rangle + i\dfrac{\rho_2}{\mu_2}\|A^{-\frac{\sigma}{2}}f^4\|^2 -i\dfrac{\rho_2}{\mu_2}\lambda^2\|A^{-\frac{\sigma}{2}}\Psi\|^2. \end{aligned}

Noting that: \|A^\frac{1-2\sigma}{2}\Phi\|^2=\|A^{-\sigma} \Phi_x\|^2, taking real part, considering \frac{1}{2}\leq\sigma\leq 1, using Cauchy-Schwarz and Young inequalities, for \varepsilon>0 exists C_\varepsilon such that

\begin{aligned} |\lambda|\|\Psi\|^2 & \leq & C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}+C_\varepsilon\|A^\frac{1-2\sigma}{2}\Phi\|^2+\varepsilon\|\Psi\|^2. \end{aligned}

As -\frac{1-2\sigma}{2} \leq 0\Longleftrightarrow \frac{1}{2}\leq \sigma\leq 1. Considering |\lambda|\geq 1, using estimative (Proof), norms \|F\|_\mathcal{H} and \|U\|_\mathcal{H}, the proof of item (ii) of this lemma is finished.

Item (iii): Finally, let’s prove the (iii) item of this lemma, performing the duality product of (?) for \dfrac{\rho_3}{K}A^{-\xi}\lambda \Theta, using (?), and recalling the self-adjointness of A^\nu, \nu \in\mathbb{R}, result in

\begin{aligned} \lambda\|\Theta\|^2 & = & -i\lambda^2\dfrac{\rho_3}{K}\|A^{-\frac{\xi}{2}}\Theta\|^2+i\dfrac{\delta}{K}\|A^\frac{1-\xi}{2}\Theta\|^2+i\dfrac{\delta}{K}\langle{A^\frac{1}{2}f^5},{A^{\frac{1}{2}-\xi}\Theta}\rangle-i\dfrac{\delta}{K}\langle{f^6},{A^{1-\xi}\theta}\rangle\\ & & +\dfrac{\gamma}{K}\langle{\sqrt{|\lambda|}\Psi},{\dfrac{\lambda}{\sqrt{|\lambda|}}A^{-\xi}\Theta_x}\rangle -i\dfrac{\gamma}{K}\langle{f^6},{A^{-\xi}\Psi_x}\rangle-i\langle{f^6},{\Theta}\rangle+i\dfrac{\rho_3}{K}\|A^{-\frac{\xi}{2}}f^6\|^2. \end{aligned}

Taking the real part, applying Cauchy-Schwarz and Young inequalities, for \varepsilon>0 exists C_\varepsilon>0 such that

\begin{gathered} |\lambda|\|\Theta\|^2\leq C\{ \|A^\frac{1-\xi}{2}\Theta\|^2+\|A^\frac{1}{2}f^5\|\|A^{\frac{1}{2}-\xi}\Theta\|\\ +\|f^6\|\|A^{1-\xi}\theta\|+\|f^6\|\|A^{-\xi}\Psi_x\|\} +C_\varepsilon |\lambda|\|\Psi\|^2+\varepsilon |\lambda|\|A^{-\xi}\Theta_x\|^2. \end{gathered}

Considering \frac{1}{2}\leq\xi\leq 1, \frac{1-2\xi}{2}\leq 0, and remembering the identities \|A^{-\xi}\Psi_x\|^2=\|A^\frac{1-2\xi}{2}\Psi\|^2, \|A^{-\xi}\Theta_x\|^2=\|A^\frac{1-2\xi}{2}\Theta\|^2 in addition to using the item (ii) of this lemma the proof of the item (iii) of this lemma is finished.

Uses Proof
Theorem

Theorem 11

VQG8

The semigroup S(t)=e^{\mathcal{B}_1t} associated to the system (?)–(?) and (?)–(?) is analytic when the three parameters \tau, \sigma and \xi vary in the interval [\frac{1}{2},1].

Proof

Proof

WK9C

We will prove this theorem will be proved using the Theorem 8, so the two conditions (8) and (8) must be proved.

From Lemma 9, the condition (8) is verified, it remains to verify the condition (8), note that proving this condition is equivalent to: let \delta>0 there exists a constant C_\delta > 0 such that the solutions of (?)–(?) for |\lambda|\geq \delta satisfy the inequality

|\lambda|\|U\|^2_\mathcal{H}\leq C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}.

It is not difficult to see that this inequality (?) arises from the inequalities of the Lemmas 4 and 10, so the proof of this theorem is finished.

Lemma

Lemma 12

N36T

Let S_2(t)=e^{\mathcal{B}_2t} be a C_0-semigroup on contractions on Hilbert space \mathcal{H}=\mathcal{H}_2, the solutions of the system (?)–(?) satisfy the inequality

\limsup_{|\lambda|\to\infty}\|(i\lambda I-\mathcal{B}_2) ^{-1}\|_{\mathcal{L}(\mathcal{H})}<\infty.
Proof of Lemma 12

To show the (3) inequality, it suffices to show that, given \delta>0 there exists a constant C_\delta>0 such that the solutions of the system (?)–(?) for |\lambda|>\delta satisfy the inequality

\dfrac{\|U\|_{\mathcal{H}}}{\|F\|_{\mathcal{H}}}\leq C_\delta\Longleftrightarrow \|U\|^2_{\mathcal{H}}=\|(i\lambda I-\mathcal{B}_2)^{-1} F\|^2_{\mathcal{H}} \leq C_\delta \|F\|_\mathcal{H}\|U\|_{\mathcal{H}}.

If F=(f^1,f^2,f^3,f^4, f^5, f^6)\in \mathcal{H} then the solution U=(\phi,\Phi,\psi, \Psi,\theta,\Theta)\in\hbox{D}(\mathcal{B}_2) the resolvent equation (i\lambda I- \mathcal{B}_2)U=F can be written in the form

\begin{aligned} i\lambda \phi-\Phi &= f^1\quad\rm{in}\quad D(A^\frac{1}{2}),\\ i\lambda \Phi-\frac{\kappa}{\rho_1}(\phi_x+\psi)_x+\dfrac{\mu}{\rho_1}\Theta_x+\frac{\mu_1}{\rho_1}A^\tau \Phi&= f^2\quad\rm{in}\quad D(A^0),\\ i\lambda \psi-\Psi &=f^3\quad \rm{in}\quad D(A^\frac{1}{2}),\\ i\lambda \Psi+\dfrac{b}{\rho_2}A\psi+\dfrac{\kappa}{\rho_2}(\phi_x+\psi)-\dfrac{\mu}{\rho_2}\Theta+\dfrac{\mu_2}{\rho_2}A^\sigma \Psi &= f^4 \quad\rm{in}\quad D(A^0),\\ i\lambda\theta-\Theta&=f^5\quad{ \rm in}\quad D(A^\frac{1}{2}),\\ i\lambda\Theta+\dfrac{\delta}{\rho_3}A\theta+\dfrac{\mu}{\rho_3}(\Phi_x+\Psi)+\dfrac{\gamma}{\rho_3}A^\xi\Theta&= f^6\quad {\rm in}\quad D(A^0). \end{aligned}

Using the fact that the operator is dissipative \mathcal{B}_2, result in

\begin{gathered} \mu_1\|A^\frac{\tau}{2}\Phi\|^2+\mu_2\|A^\frac{\sigma}{2} \Psi\|^2+\gamma\|A^\frac{\xi}{2}\Theta\|^2= \text{Re}\langle{(i\lambda -\mathcal{B}_2)U},{U}\rangle= \text{Re}\langle{F},{U}\rangle\leq \|F\|_\mathcal{H}\|U\|_\mathcal{H}. \end{gathered}

On the other hand, performing the duality product of (?) for \rho_1 \phi, and as the operators A^\nu for all \nu\in\mathbb{R} are self-adjoint, lead to

\begin{gathered} \kappa\langle{(\phi_x+\psi)},{\phi_x}\rangle= \rho_1\|\Phi\|^2+\rho_1\langle{\Phi},{f^1}\rangle -i\lambda\mu_1\|A^\frac{\tau}{2}\phi\|^2 +\mu_1\langle{A^\frac{\tau}{2} f^1},{A^\frac{\tau}{2} \phi}\rangle +\rho_1\langle{f^2},{\phi}\rangle+\mu\langle{\Theta},{\phi_x}\rangle, \end{gathered}

now performing the duality product of (?) for \rho_2 \psi, and as the operators A^\nu for all \nu\in\mathbb{R} are self-adjoint, result in

\begin{gathered} \kappa\langle{(\phi_x+\psi)},{\psi}\rangle+b\|A^\frac{1}{2}\psi\|^2=\rho_2\|\Psi\|^2 +\rho_2\langle{\Psi},{f^3}\rangle\\ +\mu\langle{\Theta},{\psi}\rangle-i\lambda\mu_2\|A^\frac{\sigma}{2}\psi\|^2+\mu_2\langle{A^\frac{\sigma}{2}f^3},{A^\frac{\sigma}{2}\psi}\rangle+\rho_2\langle{f^4},{\psi}\rangle. \end{gathered}

Adding the last 2 equations, result in

\begin{gathered} \kappa\|\phi_x+\psi\|^2+b\|A^\frac{1}{2}\psi\|^2= \rho_1\|\Phi\|^2+\rho_2\|\Psi\|^2 -i\lambda \{ \mu_1 \|A^\frac{\tau}{2}u\|^2 +\mu_2\|A^\frac{\sigma}{2}\psi\|^2 \}\\ +\rho_1\langle{\Phi},{f^1}\rangle+\rho_2\langle{\Psi},{f^3}\rangle +\mu_1\langle{A^\frac{\tau}{2}f^1},{A^\frac{\tau}{2}\phi}\rangle +\mu_2\langle{A^\frac{\sigma}{2}f^3},{A^\frac{\sigma}{2}\phi}\rangle\\ +\rho_1\langle{f^2},{\phi}\rangle+\rho_2\langle{f^4},{\psi}\rangle+\mu\langle{\Theta},{\phi_x+\psi}\rangle. \end{gathered}

Taking real part, applying Cauchy-Schwarz and Young inequalities, for \varepsilon>0 exists C_\varepsilon>0 such that

\kappa\|\phi_x+\psi\|^2+b\|A^\frac{1}{2}\psi\|^2 \leq \rho_1\|\Phi\|^2+\rho_2\|\Psi\|^2+\varepsilon\|\phi_x+\psi\|^2+ C_\varepsilon\|\Theta\|^2+C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}.

From estimative (?) and the fact 0\leq\frac{\tau}{2},\;0\leq\frac{\sigma}{2} and 0\leq\frac{\xi}{2} the continuous embedding D(A^{\theta_2}) \hookrightarrow D(A^{\theta_1}),\;\theta_2>\theta_1, result in

\begin{aligned} \kappa\|\phi_x+\psi\|^2+b\|A^\frac{1}{2}\psi\|^2&\leq& C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}. \end{aligned}

On the other hand, performing the duality product of (?) for \rho_3 \theta, and remembering that the operators A^\nu for all \nu\in\mathbb{R} are self-adjoint and from (?), lead to

\begin{gathered} \delta\|A^\frac{1}{2}\theta\|^2=\rho_3\|\Theta\|^2+\rho_3\langle{\Theta},{f^5}\rangle+\mu\langle{\Phi},{\theta_x}\rangle-\mu\langle{\Psi},{\theta}\rangle -i\lambda\gamma\|A^\frac{\xi}{2}\theta\|^2+\gamma\langle{A^\frac{\xi}{2}f^5},{A^\frac{\xi}{2}\theta}\rangle+\rho_3\langle{f^6},{\theta}\rangle. \end{gathered}

Taking the real part, using the inequalities Cauchy-Schwarz and Young, for \varepsilon>0 exists C_\varepsilon>0 which does not depend on \lambda such that

\begin{gathered} \delta\|A^\frac{1}{2}\theta\|^2 \leq C\{\|\Theta\|^2+\|\Theta\|\|f^5\|+\|A^\frac{\xi}{2}f^5\|\|A^\frac{\xi}{2}\theta\|+\|f^6\|\|\theta\|\}\\+C_\varepsilon\{\|\Phi\|^2+\|\Psi\|^2\}+\varepsilon\{ \|\theta\|^2+\|A^\frac{1}{2}\theta\|^2\}. \end{gathered}

Then, from estimative (?) and the fact 0\leq\frac{\tau}{2},\; 0\leq\frac{\sigma}{2} and 0\leq\frac{\xi}{2}\leq\frac{1}{2} the continuous embedding D(A^{\theta_2}) \hookrightarrow D(A^{\theta_1}),\;\theta_2>\theta_1, result in

\beta\kappa\delta\|A^\frac{1}{2}\theta\|^2\leq C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}.

Finally, from estimates (?), (?) and (?), the proof of this lemma is finished.

Uses Lemma 3
Lemma

Lemma 13

LNA6

Let \delta> 0. There exists a constant C_\delta > 0 such that the solutions of (?)–(?) for |\lambda|\geq \delta satisfy the inequalities

\begin{aligned} \hspace*{-1.0cm}(i)\; |\lambda|[\kappa\|\phi_x+\psi\|^2+b\|A^\frac{1}{2}\psi\|^2] &\leq & |\lambda|[\rho_1\|\Phi\|^2+\rho_2\|\Psi\|^2+C\|\Theta\|^2] +C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}.\\ \hspace*{-1.0cm}(ii)\; \delta|\lambda|\|A^\frac{1}{2}\theta\|^2 &\leq & C |\lambda|\{ \|\Theta\|^2+\|\Phi\|^2\}+ C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}. \end{aligned}
Used by Proof, Proof
Proof of Lemma 13

Item (i): Performing the duality product of (?) for \rho_1\lambda \phi, and remembering that the operators A^\nu for all \nu\in \mathbb{R} are self-adjunct, lead to

\begin{aligned} \kappa\lambda\langle{(\phi_x+\psi)},{\phi_x}\rangle& =&\rho_1\lambda\|\Phi\|^2+\rho_1\langle{\lambda \Phi},{f^1}\rangle -i\mu_1\|A^\frac{\tau}{2}\Phi\|^2\\ & & -i\mu_1\beta\gamma\langle{A^\tau \Phi},{f^1}\rangle +i\rho_1\langle{f^2},{\Phi}\rangle+i\rho_1\langle{f^2},{f^1}\rangle+\mu\langle{\Theta},{\lambda\phi_x}\rangle. \end{aligned}

Now performing the duality product of (?) for \rho_2\lambda \psi, and remembering that the operators A^\nu for all \nu\in \mathbb{R} are self-adjunct, resulting in

\begin{aligned} \lambda[\kappa\langle{(\phi_x+\psi)},{\psi}\rangle+ b\|A^\frac{1}{2}\psi\|^2]\hspace*{-0.3cm} & = &\hspace*{-0.3cm}\rho_2\lambda\|\Psi\|^2+\rho_2\langle{\lambda \Psi},{f^3}\rangle- i\mu_2\|A^\frac{\sigma}{2}\Psi\|^2 -i\mu_2\langle{A^\sigma \Psi},{f^3}\rangle \\ & & +i\rho_2\langle{f^4},{\Psi}\rangle +i\rho_2\langle{f^4},{f^3}\rangle+\mu\langle{\Theta},{\lambda\psi}\rangle. \end{aligned}

Adding the last 2 equations, result in

\begin{gathered} \lambda[\kappa\|\phi_x+\psi\|^2+b\|A^\frac{1}{2}\psi\|^2] = \lambda[\rho_1\|\Phi\|^2+\rho_2\|\Psi\|^2] -i\{\mu_1\|A^\frac{\tau}{2}\Phi\|^2\\+\mu_2\|A^\frac{\sigma}{2}\Psi\|^2\}+\rho_1\langle{\lambda \Phi},{f^1}\rangle+\rho_2\langle{\lambda \Psi},{f^3}\rangle-i\{\mu_1 \langle{A^\tau \Phi},{f^1}\rangle\\ +\mu_2\langle{A^\sigma \Psi},{f^3}\rangle \} +i\rho_1\langle{f^2},{\Phi}\rangle +i\rho_2\langle{f^4},{\Psi}\rangle +i\rho_1\langle{f^2},{f^1}\rangle\\ +i\rho_2\langle{f^4},{f^3}\rangle+\mu\langle{\sqrt{|\lambda|}\Theta},{\dfrac{\lambda}{\sqrt{|\lambda|}}(\phi_x+\psi)}\rangle. \end{gathered}

On the other hand, from (?) and (?), result in

\begin{gathered} \rho_1\langle{\lambda \Phi},{f^1}\rangle+\rho_2\langle{\lambda \Psi},{f^3}\rangle =i\{ \kappa\langle{(\phi_x+\psi)},{f^1_x}\rangle-\mu\langle{\Theta},{f^1_x}\rangle\\ +\mu_1\langle{A^\tau \Phi},{f^1}\rangle-\rho_1\langle{f^2},{f^1}\rangle +b\langle{A^\frac{1}{2}\psi},{A^\frac{1}{2}f^3}\rangle+\kappa\langle{(\phi_x+\psi)},{f^3}\rangle\\ -\mu\langle{\Theta},{f^3}\rangle+\mu_2\langle{A^\sigma \Psi},{f^3}\rangle-\rho_2\langle{f^4},{f^3}\rangle\}. \end{gathered}

Using the identity (?) in the (?) equation and simplifying, leads to

\begin{aligned} \hspace*{-0.45cm}\lambda[\kappa\|\phi_x+\psi\|^2+b\|A^\frac{1}{2}\psi\|^2] \hspace*{-0.25cm} &= & \hspace*{-0.25cm} \lambda[\rho_1\|\Phi\|^2+\rho_2\|\Psi\|^2] -i\{\mu_1\|A^\frac{\tau}{2}\Phi\|^2 +\mu_2\|A^\frac{\sigma}{2}\Psi\|^2\}+i\kappa\langle{\phi_x},{f_x^1}\rangle\\ & & \hspace*{-0.25cm}+i\kappa\langle{\psi},{f_x^1}\rangle -i\mu\langle{\Theta},{f^1_x}\rangle +ib \langle{A^\frac{1}{2}\psi},{A^\frac{1}{2}f^3}\rangle+i\kappa\langle{\phi_x},{f^3}\rangle+i\kappa\langle{\psi},{f^3}\rangle\\ & &\hspace*{-0.25cm} -i\mu\langle{\Theta},{f^3}\rangle+i\rho_1\langle{f^2},{\Phi}\rangle+i\rho_2\langle{f^4},{\Psi}\rangle +\mu\langle{\sqrt{|\lambda|}\Theta},{\dfrac{\lambda}{\sqrt{|\lambda|}}(\phi_x+\psi)}\rangle. \end{aligned}

As for \varepsilon>0, exists C_\varepsilon>0, such that \bigg|\mu\langle{\sqrt{|\lambda|}\Theta},{\dfrac{\lambda}{\sqrt{|\lambda|}}(\phi_x+\psi)}\rangle\bigg|\leq C_\varepsilon|\lambda|\|\Theta\|^2+\varepsilon|\lambda|\|\phi_x+\psi\|^2, taking the real part of the equation (?), applying the Cauchy-Schwarz and Young inequalities and estimative (Proof) of Lemma 12, the proof of item (i) this lemma is completed.

On the other hand, performing the duality product of (?) for \rho_3\lambda \theta, and remembering that the operators A^\nu for all \nu\in \mathbb{R } are self-adjunct, result in

\delta\lambda\|A^\frac{1}{2}\theta\|^2=\rho_3\lambda\| \Theta\|^2+i\mu\langle{\Phi},{\Theta_x}\rangle-i\gamma\|A^\frac{\xi}{2}\Theta\|^2+i\rho_3\langle{f^6},{\Theta}\rangle-i\mu\langle{\Psi},{\Theta}\rangle.

As

i\mu\langle{\Phi},{\Theta_x}\rangle=i\mu\langle{\Phi},{i\lambda\theta_x-f_x^5}\rangle=\mu\langle{\sqrt{|\lambda|}\Phi},{\dfrac{\lambda}{\sqrt{|\lambda|}}\theta_x}\rangle -i\mu\langle{\Phi},{f_x^5}\rangle.

Then, using (?) in (?) and taking real part, lead to

\begin{gathered} \delta|\lambda|\|A^\frac{1}{2}\theta\|^2=\rho_3\lambda\| \Theta\|^2+\mu\langle{\sqrt{|\lambda|}\Phi},{\dfrac{\lambda}{\sqrt{|\lambda|}}\theta_x}\rangle -i\mu\langle{\Phi},{f_x^5}\rangle +i\rho_3\langle{f^6},{\Theta}\rangle-i\mu\langle{\Psi},{\Theta}\rangle, \end{gathered}

applying Cauchy-Schwarz and Young inequalities, for \varepsilon>0 exists C_\varepsilon>0, such that

\delta|\lambda|\|A^\frac{1}{2}\theta\|^2 \leq \rho_3|\lambda|\|\Theta\|^2+ \varepsilon|\lambda|\|A^\frac{1}{2}\theta\|^2+ C_\varepsilon|\lambda|\|\Phi\|^2 +C\{\|\Phi\|\|A^\frac{1}{2}f^5\|+\|f^6\|\|\Theta\|+\|\Phi\|^2+\|\Theta\|^2 \}.

Then, from Lemma 12, the proof of item (ii) of this lemma is finished.

Theorem

Theorem 14

SZ4J

The semigroup S(t)=e^{\mathcal{B}_2t} associated to system (?)–(?) is of Gevrey class s>\dfrac{1+r}{2r} for r=\min\{\tau,\sigma,\xi\}, for all (\tau,\sigma,\xi)\in R_{CG}:=(0,1)^3 and \frac{2r}{r+1}\in (0,1).

Proof of Theorem 14

From the resolvent equation F=(i\lambda I-\mathcal{B}_2)U for \lambda\in\mathbb{R}, result in U=(i\lambda I-\mathcal{B}_2)^{-1}F. Furthermore to show (6) this is theorem(6) it is enough to show:

|\lambda |^\frac{2r}{r+1}\|U\|_\mathcal{H}^2\leq C_\delta \|F\|_\mathcal{H}\|U\|_\mathcal{H}\quad\rm{for}\quad \frac{2r}{r+1} \in (0,1).

where r=\min\{\tau,\sigma,\xi\}, for all (\tau,\sigma,\xi)\in (0,1)^3.

Next, |\lambda|^\frac{2\tau}{1+\tau}\|\Phi\|, |\lambda|^\frac{2\sigma}{1+\sigma}\|\Psi\| and |\lambda|^\frac{2\xi}{1+\xi}\|\Theta\| will be estimated.

Let’s start by estimating the term |\lambda|^\frac{2\tau}{1+\tau}\|\Phi\|: It is assumed that |\lambda|>1. Set \Phi=\Phi_1+\Phi_2, where \Phi_1\in D(A) and \Phi_2\in D(A^0), with

\begin{gathered} i\lambda \Phi_1+A \Phi_1=f^2, \hspace{1cm} i\lambda \Phi_2=-\dfrac{\kappa}{\rho_1}A\phi+\dfrac{\kappa}{\rho_1}\psi_x -\dfrac{\mu}{\rho_1}\Theta_x-\dfrac{\mu_1}{\rho_1}A^\tau \Phi+A\Phi_1. \end{gathered}

Firstly, applying the product duality on the first equation in (?) by \Phi_1, then by A^{\Phi_1} and recalling that the operator A is self-adjoint, resulting in self-adjoint, resulting in

|\lambda|\|\Phi_1\| +|\lambda|^\frac{1}{2}\|A^\frac{1}{2}\Phi_1\|+\|A\Phi_1\|\leq C\|F\|_\mathcal{H}.

Applying the A^{-\frac{1}{2}} operator on the second equation of (?), lead to

i\lambda A^{-\frac{1}{2}}\Phi_2= -\dfrac{\kappa}{\rho_1}A^\frac{1}{2}\phi+\dfrac{\kappa}{\rho_1}A^{-\frac{1}{2}}\psi_x-\dfrac{\mu}{\rho_1}A^{-\frac{1}{2}}\Theta_x-\dfrac{\mu_1}{\rho_1}A^{\tau-\frac{1}{2}} \Phi+A^\frac{1}{2}\Phi_1,

then, as \|A^{-\frac{1}{2}}\psi_x\|^2=\|\psi\|^2 and \tau-\frac{1}{2}\leq \frac{\tau}{2} taking into account the continuous embedding D(A^{\theta_2}) \hookrightarrow D(A^{\theta_1}),\;\theta_2>\theta_1, result in

|\lambda|^2\|A^{-\frac{1}{2}} \Phi_2\|^2 \leq C\{\|A^\frac{1}{2}\phi\|^2+ \|A^\frac{\tau}{2}\Phi\|^2+\|\psi\|^2+\|\Theta\|^2\}+\|A^\frac{1}{2}\Phi_1\|^2

Using (Proof) and estimative (?), result in

\|A^{-\frac{1}{2}}\Phi_2\|^2\leq C|\lambda|^{-\frac{4\tau+2}{\tau+1}}\{ |\lambda|^\frac{2\tau}{\tau+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}\quad {\rm for}\quad 0\leq\tau\leq 1.

On the other hand, from \Phi_2=\Phi-\Phi_1, (?) and as \frac{\tau}{2}\leq\frac{1}{2} the second inequality of (?), result in

\begin{aligned} \|A^\frac{\tau}{2} \Phi_2\|^2& \leq &C\{ \|A^\frac{\tau}{2} \Phi\|^2+\|A^\frac{\tau}{2}\Phi_1\|^2\} \leq C|\lambda|^{-\frac{2\tau}{\tau+1}}\{|\lambda|^\frac{2\tau}{\tau+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}. \end{aligned}

Now, by Lions’ interpolations inequality 0\in[-\frac{1}{2},\frac{\tau}{2}], result in

\|\Phi_2\|^2\leq C (\|A^{-\frac{1}{2}}\Phi_2\|^2)^\frac{\tau}{1+\tau}(\|A^\frac{\tau}{2}\Phi_2\|^2)^\frac{1}{1+\tau}.

Then, using (?) and (?) in (?), result in

\|\Phi_2\|^2\leq C|\lambda|^{\frac{-4\tau}{(1+\tau)}}\{ |\lambda|^\frac{2\tau}{\tau+1} \|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}.

Therefore, as \|\Phi\|^2\leq \|\Phi_1\|^2+ \|\Phi_2\|^2 from first inequality of (?), (?) and |\lambda|^{-2}\leq |\lambda|^\frac{-4\tau}{1+\tau}, we have

\begin{gathered} |\lambda|^\frac{2\tau}{1+\tau}\|\Phi\|\leq C_\delta\|F\|_\mathcal{H}\\ \Longleftrightarrow \quad |\lambda|\|\Phi\|^2\leq C_\delta|\lambda|^\frac{1-3\tau}{1+\tau}\{|\lambda|^\frac{2\tau}{\tau+1} \|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}\quad\rm{for}\quad 0\leq\tau\leq 1. \end{gathered}

On the other hand, let’s now estimate the missing term |\lambda|^\frac{2\sigma}{1+\sigma}\|\Psi\|: Set \Psi=\Psi_1+\Psi_2, where \Psi_1\in D(A) and \Psi_2\in D(A^0), with

\hspace*{-0.2cm}i\lambda \Psi_1+A \Psi_1=f^4 \quad{\rm and} \quad i\lambda \Psi_2=-\dfrac{b}{\rho_2}A\psi -\dfrac{\kappa}{\rho_2}\phi_x-\dfrac{\kappa}{\rho_2}\psi+\dfrac{\mu}{\rho_2}\Theta-\dfrac{\mu_2}{\rho_2}A^\sigma \Psi+A\Psi_1.

Firstly, applying the product duality on the first equation in (?) by \Psi_1, then by A\Psi_1 and recalling that the operator A is self-adjoint, resulting in

|\lambda|\|\Psi_1\| +|\lambda|^\frac{1}{2}\|A^\frac{1}{2}\Psi_1\|+\|A\Psi_1\|\leq C \|F\|_\mathcal{H}.

It arises from the second equation in (?) that

i\lambda A^{-\frac{1}{2}}\Psi_2= -\dfrac{b}{\rho_2}A^\frac{1}{2}\psi-\dfrac{\kappa}{\rho_2}A^{-\frac{1}{2}}\phi_x-\dfrac{\kappa}{\rho_2}A^{-\frac{1}{2}}\psi+\dfrac{\mu}{\rho_2}A^{-\frac{1}{2}}\Theta-\dfrac{\mu_2}{\rho_2}A^{\sigma-\frac{1}{2}}\Psi+A^\frac{1}{2}\Psi_1,

then, as \|A^{-\frac{1}{2}}\phi_x\|^2=\|\phi\|^2, \|A^{-\frac{1}{2}}\Theta_x\|^2=\|\Theta\|^2 and \sigma-\frac{1}{2}\leq \frac{\sigma}{2} taking into account the continuous embedding D(A^{\theta_2}) \hookrightarrow D(A^{\theta_1}),\;\theta_2>\theta_1, result in

|\lambda|^2\|A^{-\frac{1}{2}} \Psi_2\|^2\leq C\{ \|\phi\|^2+\|A^{\frac{1}{2}}\psi\|^2+\|A^{-\frac{1}{2}}\Theta\|^2+\|A^\frac{\sigma}{2}\Psi\|^2\}+\|A^\frac{1}{2}\Psi_1\|^2

Using estimative (Proof) of Lemma 12 and second estimative of (?), yields

\|A^{-\frac{1}{2}}\Psi_2\|^2\leq C|\lambda|^{-\frac{4\sigma+2}{\sigma+1}}\{|\lambda|^\frac{2\sigma}{\sigma+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|_\mathcal{H}\}\quad\text{for}\quad 0\leq \sigma\leq 1.

On the other hand, from \Psi_2=\Psi-\Psi_1, (?) and as 0\leq\frac{\sigma}{2}\leq\frac{1}{2}, the second inequality of (?), result in

\begin{gathered} \|A^\frac{\sigma}{2} \Psi_2\|^2 \leq C\{ \|A^\frac{\sigma}{2} \Psi\|^2+\|A^\frac{\sigma}{2}\Psi_1\|^2\} \leq C|\lambda|^{-\frac{2\sigma}{\sigma+1}}\{|\lambda|^\frac{2\sigma}{\sigma+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}. \end{gathered}

Now using interpolation inequality 0\in [-\frac{1}{2},\frac{\sigma}{2}]. Since

0=\eta\bigg(-\dfrac{1}{2}\bigg)+(1-\eta)\dfrac{\sigma}{2},\quad {\rm for}\quad \eta=\dfrac{\sigma}{1+\sigma},

using (?) and (?) we get that

\begin{aligned} \|\Psi_2\|^2 &\leq & C_\delta (\|A^{-\frac{1}{2}}\Psi_2\|^2)^{\eta} (\|A^\frac{\sigma}{2}\Psi_2\|^2)^{1-\eta} \leq C |\lambda|^{\frac{-4\sigma}{1+\sigma}}\{|\lambda|^\frac{2\sigma}{\sigma+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}. \end{aligned}

Also, as \|\Psi\|^2\leq C\{\|\Psi_1\|^2+ \|\Psi_2\|^2\} from first inequality of (?), estimative (?) and as |\lambda|^{-2}\leq |\lambda|^\frac{-2\sigma}{1+\sigma}, result in

\begin{gathered} |\lambda|^\frac{2\sigma}{1+\sigma}\|\Psi\|\leq C_\delta\|F\|_\mathcal{H}\\ \Longleftrightarrow\; |\lambda|\|\Psi\|^2 \leq C_\delta|\lambda|^\frac{1-3\sigma}{\sigma+1}\{|\lambda|^\frac{2\sigma}{\sigma+1} \|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}\quad\rm{for}\quad 0\leq\sigma\leq 1. \end{gathered}

Finally, let’s now estimate the missing term |\lambda|^\frac{2\xi}{1+\xi}\|\Theta\|. Set \Theta=\Theta_1+\Theta_2, where \Theta_1\in D(A) and \Theta_2\in D(A^0), with

\begin{gathered} i\lambda \Theta_1+A \Theta_1=f^6 \quad{\rm and}\quad i\lambda \Theta_2=-\dfrac{\delta}{\rho_3}A\theta -\dfrac{\mu}{\rho_3}\Phi_x-\dfrac{\mu}{\rho_3}\Psi-\dfrac{\gamma}{\rho_3}A^\xi \Theta+A\Theta_1. \end{gathered}

Firstly, applying the product duality the first equation in (?) by \Theta_1, then by A\Theta_1 and as the operators A^\nu for all \nu\in\mathbb{R} are self-adjoint, result in

|\lambda|^2\|\Theta_1\|^2+|\lambda|\|A^\frac{1}{2}\Theta_1\|^2+\|A\Theta_1\|^2\leq C \|F\|^2_\mathcal{H}.

It arises from the second equation in (?) that

i\lambda A^{-\frac{1}{2}}\Theta_2= -\dfrac{\delta}{\rho_3}A^\frac{1}{2}\theta -\dfrac{\mu}{\rho_3}A^{-\frac{1}{2}}\Phi_x-\dfrac{\mu}{\rho_3}A^{-\frac{1}{2}}\Psi-\dfrac{\gamma}{\rho_3}A^{\xi-\frac{1}{2}} \Theta+A^\frac{1}{2}\Theta_1,

then, as \|A^{-\frac{1}{2}}\Phi_x\|^2=\|\Phi\|^2, -\frac{1}{2}\leq 0 and \xi-\frac{1}{2}\leq \frac{\xi}{2} taking into account the continuous embedding D(A^{\theta_2}) \hookrightarrow D(A^{\theta_1}),\;\theta_2>\theta_1 and using estimative (Proof) and estimative (?), lead to

\|A^{-\frac{1}{2}}\Theta_2\|^2\leq C |\lambda|^{-\frac{4\xi+2}{\xi+1}}\{|\lambda|^\frac{2\xi}{\xi+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|_\mathcal{H}\}.

On the other hand, from \Theta_2=\Theta-\Theta_1, (?) and as \frac{\xi}{2}\leq\frac{1}{2} the second inequality of (?), result in

\begin{aligned} \|A^\frac{\xi}{2} \Theta_2\|^2 & \leq & C|\lambda|^{-\frac{2\xi}{\xi+1}}\{|\lambda|^\frac{2\xi}{\xi+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\}. \end{aligned}

Now using interpolation inequality 0\in [-\frac{1}{2},\frac{\xi}{2}]. Since

0=\eta\bigg(-\dfrac{1}{2}\bigg)+(1-\eta)\dfrac{\xi}{2},\quad {\rm for}\quad \eta=\dfrac{\xi}{1+\xi},

using (?) and (?) we get that

\begin{aligned} \|\Theta_2\|^2 &\leq & C_\delta (\|A^{-\frac{1}{2}}\Theta_2\|^2)^{\eta} (\|A^\frac{\xi}{2}\Theta_2\|^2)^{1-\eta} \leq |\lambda|^{\frac{-2\xi}{\xi+1}}\{\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} \end{aligned}

Also as, \|\Theta\|^2\leq C\{ \|\Theta_1\|^2+ \|\Theta_2\|^2\} from first inequality of (?), (?) and as |\lambda|^{-2}\leq |\lambda|^\frac{-4\xi}{1+\xi}, result in

\begin{gathered} |\lambda|^\frac{2\xi}{1+\xi}\|\Theta\|\leq C_\delta\|F\|_\mathcal{H}\\ \Longleftrightarrow\quad |\lambda|\|\Theta\|^2\leq C_\delta |\lambda|^\frac{1-3\xi}{1+\xi}\{|\lambda|^\frac{2\xi}{\xi+1} \|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F||^2_\mathcal{H}\}\quad\rm{for}\quad 0\leq\xi\leq 1. \end{gathered}

Finally, using (?), (?) and (?) in the inequality item (i) of Lemma 13, result in

\begin{gathered} \beta\gamma[\kappa\|\phi_x+\psi\|^2+b\|A^\frac{1}{2}\psi\|^2] \leq C\big\{ |\lambda|^\frac{-2\tau}{1+\tau}+|\lambda|^\frac{-2\sigma}{1+\sigma}+|\lambda|^\frac{-2\xi}{1+\xi}+|\lambda|^{-1}\big\} \|F\|_\mathcal{H}\|U\|_\mathcal{H}\\ +C\{ |\lambda|^{-\frac{4\tau}{\tau+1}}+ |\lambda|^{-\frac{4\sigma}{\sigma+1}}+ |\lambda|^{-\frac{4\xi}{\xi+1}}\}\|F\|^2_\mathcal{H}, \end{gathered}

then

\begin{gathered} \beta\gamma|\lambda|[\kappa\|\phi_x+\psi\|^2+b\|A^\frac{1}{2}\psi\|^2]\\ \leq C_\delta\left\{ \begin{array}{ccc} |\lambda|^\frac{1-3\tau}{1+\tau}\{ |\lambda|^\frac{2\tau}{\tau+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} &\text{for} & \tau\leq \sigma\quad{\rm and}\quad \tau\leq \xi, \\\\ |\lambda|^\frac{1-3\sigma}{1+\sigma}\{ |\lambda|^\frac{2\sigma}{\sigma+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} &\text{for} & \sigma \leq \tau\quad{\rm and}\quad \sigma\leq \xi,\\\\ |\lambda|^\frac{1-3\xi}{1+\xi}\{ |\lambda|^\frac{2\xi}{\xi+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} &\text{for} & \xi\leq \sigma\quad{\rm and}\quad \xi\leq \tau. \end{array}\right. \end{gathered}

Analogously, using (?) and (?) in the inequality item (ii) of Lemma 13, result in

\delta|\lambda| \|A^\frac{1}{2}\theta\|^2 \leq C_\delta \left\{ \begin{array}{ccc} |\lambda|^\frac{1-3\tau}{1+\tau}\{ |\lambda|^\frac{2\tau}{\tau+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} &\text{for} & \tau\leq \xi, \\\\ |\lambda|^\frac{1-3\xi}{1+\xi}\{ |\lambda|^\frac{2\xi}{\xi+1}\|F\|_\mathcal{H}\|U\|_\mathcal{H}+\|F\|^2_\mathcal{H}\} &\text{for} & \xi\leq \tau. \end{array}\right.

Finally, summing the the estimates (?), (?), (?), (?) and (?) and applying Young inequality, we obtain

\left\{ \begin{array}{ccc} |\lambda|^\frac{2\tau}{\tau+1}\|U\|_\mathcal{H}\leq C_\delta|\|F\|_\mathcal{H} &\text{for} & \tau\leq \sigma\quad{\rm and}\quad \tau\leq \xi, \\\\ |\lambda|^\frac{2\sigma}{1+\sigma}\|U\|_\mathcal{H}\leq C_\delta\|F\|_\mathcal{H} &\text{for} & \sigma \leq \tau\quad{\rm and}\quad \sigma\leq \xi,\\\\ |\lambda|^\frac{2\xi}{1+\xi}\|U\|_\mathcal{H}\leq C_\delta \|F\|_\mathcal{H} & \text{for} & \xi\leq \sigma\quad{\rm and}\quad \xi\leq \tau. \end{array}\right.

for (\tau,\sigma, \xi)\in (0,1)^3 the proof of this Theorem 14, is finished.

Lemma

Lemma 15

N0QN

Let \delta> 0. There exists a constant C_\delta > 0 such that the solutions of (?)–(?) for |\lambda|\geq \delta satisfy the inequality

\begin{aligned} (i)\quad |\lambda|[\|\Phi\|^2 +\|\Theta\|^2]\leq C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}\qquad{\rm for}\qquad \frac{1}{2}\leq\tau=\xi\leq 1. \\ (ii)\quad |\lambda|\|\Psi\|^2\leq C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}\qquad{\rm for}\qquad \frac{1}{2}\leq\sigma\leq 1. \end{aligned}
Used by Proof
Proof of Lemma 15

Item (i): Realizing the duality product of (?) with \dfrac{\rho_1}{\gamma}A^{-\tau}\lambda \Phi and using the property that the operator A^\nu is self-adjoint for all \nu\in \mathbb{R} and using the equation (?), result in

\begin{aligned} \dfrac{\mu_1}{\gamma}\lambda\|\Phi\|^2 &= i\dfrac{\kappa}{\gamma}\|A^\frac{1-\tau}{2} \Phi\|^2+i\dfrac{\kappa}{\gamma}\langle{A^\frac{1}{2}f^1},{A^{\frac{1}{2}-\tau}\Phi}\rangle+i\dfrac{\kappa}{\gamma}\langle{\Psi},{A^{-\tau}\Phi_x}\rangle-i\dfrac{\kappa}{\mu_1}\langle{f_x^3},{A^{-\tau}\Phi}\rangle\\ & &-i\langle{f^2},{\Phi}\rangle -i\dfrac{\kappa}{\gamma}\langle{f^2},{A^{1-\tau}\phi}\rangle +i\dfrac{\kappa}{\gamma}\langle{A^{-\tau}f^2},{\psi_x}\rangle+i\dfrac{\rho_1}{\gamma}\|A^{-\frac{\tau}{2}}f^2\|^2 \\ & & -i\dfrac{\rho_1}{\gamma}\lambda^2\|A^{-\frac{\theta}{2}}\Phi\|^2 +\dfrac{\mu}{\gamma}\langle{\Theta},{\lambda A^{-\tau}\Phi_x}\rangle. \end{aligned}

On the other hand, now performing the duality product of (?) for \dfrac{\rho_3}{\gamma}A^{-\xi}\lambda \Theta, using (?), and as the operators A^\nu for all \nu\in\mathbb{R} are self-adjoint, result in

\begin{aligned} \lambda\|\Theta\|^2 \hspace*{-0.3cm}& = &\hspace*{-0.3cm} -i\lambda^2\dfrac{\rho_3}{\gamma}\|A^{-\frac{\xi}{2}}\Theta\|^2+i\dfrac{\delta}{\gamma}\|A^\frac{1-\xi}{2}\Theta\|^2+i\dfrac{\delta}{\gamma}\langle{A^\frac{1}{2}f^5},{A^{\frac{1}{2}-\xi}\Theta}\rangle-i\dfrac{\delta}{\gamma}\langle{f^6},{A^{1-\xi}\theta}\rangle-i\dfrac{\mu}{\gamma}\langle{f^6},{A^{-\xi}\Phi_x}\rangle\\ & & -i\dfrac{\mu}{\gamma}\langle{f^6},{A^{-\xi}\Psi}\rangle -i\langle{f^6},{\Theta}\rangle+i\dfrac{\rho_3}{\gamma}\|A^{-\frac{\xi}{2}}f^6\|^2-\dfrac{\mu}{\gamma}\langle{\lambda A^{-\xi}\Phi_x},{\Theta}\rangle. \end{aligned}

Imposing the condition \tau=\xi and since {\rm Re}\{\langle{\Theta},{\lambda A^{-\tau}\Phi_x}\rangle-\langle{\lambda A^{-\tau}\Phi_x},{\Theta}\rangle\}=0, adding the equations (Proof) and (?) and then taking the real part, result in

\begin{aligned} \lambda\bigg [ \dfrac{\mu_1}{\gamma}\|\Phi\|^2+\|\Theta\|^2\bigg] &= & i\dfrac{\kappa}{\gamma}\langle{A^\frac{1}{2}f^1},{A^{\frac{1}{2}-\tau}\Phi}\rangle+i\dfrac{\kappa}{\gamma}\langle{\Psi},{A^{-\tau}\Phi_x}\rangle-i\dfrac{\kappa}{\mu_1}\langle{f_x^3},{A^{-\tau}\Phi}\rangle-i\langle{f^2},{\Phi}\rangle\\ & & -i\dfrac{\kappa}{\gamma}\langle{f^2},{A^{1-\tau}\phi}\rangle +i\dfrac{\kappa}{\gamma}\langle{A^{-\tau}f^2},{\psi_x}\rangle +i\dfrac{\delta}{\gamma}\langle{A^\frac{1}{2}f^5},{A^{\frac{1}{2}-\xi}\Theta}\rangle \\ & & -i\dfrac{\delta}{\gamma}\langle{f^6},{A^{1-\xi}\theta}\rangle-i\dfrac{\mu}{\gamma}\langle{f^6},{A^{-\xi}\Phi_x}\rangle -i\dfrac{\mu}{\gamma}\langle{f^6},{A^{-\xi}\Psi}\rangle -i\langle{f^6},{\Theta}\rangle \end{aligned}

From |i\frac{\kappa}{\gamma}\langle{\Psi},{A^{-\tau}\Psi_x}\rangle|\leq C\{\|\Psi\|^2+\|A^\frac{1-2\tau}{2}\Phi\|^2\} and \frac{1}{2}\leq\tau=\xi\leq 1, applying Cauchy-Schwarz and Young inequalities, norms \|F\|_\mathcal{H} and \|U\|_\mathcal{H}, the proof of item (i) of this lemma is finished.

Item (ii): Similarly, performing the duality product of (?) for \dfrac{\rho_2}{\mu_2}A^{-\sigma}\lambda \Psi, using (?), and recalling the self-adjointness of A^\nu, \nu \in\mathbb{R}, lead to

\begin{aligned} \lambda\|\Psi\|^2 \hspace*{-0.2cm}&=\hspace*{-0.2cm} -i\dfrac{\rho_2}{\mu_2}\lambda^2\|A^{-\frac{\sigma}{2}}\Psi\|^2+i\dfrac{b}{\mu_2}\|A^\frac{1-\sigma}{2}\Psi\|^2+i\dfrac{b}{\mu_2}\langle{A^\frac{1}{2}f^3},{A^{\frac{1}{2}-\sigma}\Psi}\rangle-i\dfrac{\kappa}{\mu_2}\langle{\Phi},{A^{-\sigma}\Psi_x}\rangle\\ & & +i\dfrac{\kappa}{\mu_2}\langle{f^1_x},{A^{-\sigma}\Psi}\rangle+i\dfrac{\kappa}{\mu_2}\|A^{-\frac{\sigma}{2}}\Psi\|^2+i\dfrac{\kappa}{\mu_2}\langle{f^3},{A^{-\sigma}\Psi}\rangle \\ & &+\dfrac{\mu}{\mu_2}\langle{\sqrt{|\lambda|}\Theta},{\dfrac{\lambda}{\sqrt{|\lambda|}}A^{-\sigma}\Psi}\rangle-i\dfrac{b}{\mu_2}\langle{f^4},{A^{1-\sigma}\Psi}\rangle-i\dfrac{\kappa}{\mu_2}\langle{f^4},{A^{-\sigma}\phi_x}\rangle\\ & &-i\dfrac{\kappa}{\mu_2}\langle{f^4},{A^{-\sigma}\psi}\rangle +i\dfrac{\mu}{\mu_2}\langle{f^4},{A^{-\sigma}\Theta}\rangle-i\langle{f^4},{\Psi}\rangle+i\dfrac{\rho_2}{\mu_2}\|f^4\|^2. \end{aligned}

Noting that: For \frac{1}{2}\leq\sigma\leq 1 lead to \frac{1-\sigma}{2}\leq \frac{\sigma}{2}, \frac{1}{2}-\sigma\leq 0 and \frac{1-2\sigma}{2}\leq 0, on the other hand as \|A^{-\sigma} \Psi_x\|^2=\|A^\frac{1-2\sigma}{2}\Psi\|^2, taking real part and considering that \frac{1}{2}\leq\sigma\leq 1 and using Cauchy-Schwarz and Young inequalities, norms \|F\|_\mathcal{H} and \|U\|_\mathcal{H}, for \varepsilon>0, exists C_\varepsilon>0 independent of \lambda such that

\begin{aligned} |\lambda|\|\Psi\|^2 & \leq & C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}+C\{\|\Phi\|^2+\|\Psi\|^2\}+C_\varepsilon|\lambda|\|\Theta\|^2+\varepsilon|\lambda|\|\Psi\|^2. \end{aligned}

From (Proof) and item (i) of this lemma, the proof of item (ii) of this lemma is finished.

Uses Proof
Theorem

Theorem 16

K6UH

The semigroup S(t)=e^{\mathcal{B}_2t} associated to the system (?)–(?) is analytic when the three parameters \tau, \sigma and \xi vary in the interval [\frac{1}{2},1] with \tau=\xi:

80-35

The paper has a picture here, which this page does not show.

FIG. 02: Region R_{A2} of Analyticity de S(t)=e^{\mathcal{B}_2t}

Proof of Theorem 16

This theorem will be proved again using now the Theorem 8, so the two conditions (8) and (8) must be proved.

Here the condition’s test (8) will be omitted because it is completely similar to the test already conducted for the first the first system.

Next, the condition (8) is proved, note that proving this condition is equivalent to show, let \delta>0. There exists a constant C_\delta > 0 such that the solutions of (?)–(?) and (?)–(?) for |\lambda|\geq \delta satisfy the inequality

|\lambda|\|U\|^2_\mathcal{H}\leq C_\delta\|F\|_\mathcal{H}\|U\|_\mathcal{H}.

It is not difficult to see that this inequality (?) follows from the inequalities of the Lemmas 13 and 15, so the proof of this theorem is finished.

Remark

Remark 17 (Asymptotic Behavior)

W56L

It is emphasized that thanks to the Lemmas 3 and 9 the exponential decay of the semigroup S_1(t)=e^{\mathcal{B}_1t} is obtained when the 3 parameters take values in the closed interval [0,1], also for the second system using the Lemmas 12 and proving in a completely similar way to the proof of Lemma 9 which i\mathbb{R}\subset\rho(\mathcal{B}_2) the corresponding semigroup S_2(t)=e^{\mathcal{B}_2t}, will be exponentially stable when the 3 parameters take values in the closed interval [0, 1 ].

Source and licence

The text on this page is from arXiv:2211.10816v2 by Filomena Barbosa Rodrigues Mendes, Lesly Daiana Barbosa Sobrado and Fredy Maglorio Sobrado Suárez, under the licence CC BY 4.0. The corpus made these changes: it read the items out of the LaTeX source, it expanded the macros of the paper, it put each proof after its statement, and it converted the text to HTML. The formulas are the TeX of the paper, and KaTeX renders them. A formula or a reference can look different from the PDF. The page is published under CC BY 4.0.

The record of each item is in analysis/objects/2211/2211.10816.jsonl.