Skip to main content
Log in

Stackelberg–Nash null controllability for some linear and semilinear degenerate parabolic equations

  • Original Article
  • Published:
Mathematics of Control, Signals, and Systems Aims and scope Submit manuscript

Abstract

This paper deals with the application of Stackelberg–Nash strategies to the null controllability of degenerate parabolic equations. We assume that we can act on the system through a hierarchy of controls. A first control (the leader) is assumed to determine the policy; then, a Nash equilibrium pair (corresponding to a noncooperative multi-objective optimization strategy) is found; this governs the action of other controls (the followers). This way, the state of the system is driven to zero and, consequently, we solve a hierarchical null controllability problem. The main novelty in this paper is that the physical systems are governed by linear or semilinear 1D heat equations with degenerate coefficients.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

References

  1. Alabau-Boussouira F, Cannarsa P, Fragnelli G (2006) Carleman estimates for degenerate parabolic operators with applications to null controllability. J Evol Equ 6(2):161–206

    Article  MathSciNet  Google Scholar 

  2. Araruna FD, Fernández-Cara E, Guerrero S, Santos MC (2017) New results on the Stackelberg–Nash exact control of linear parabolic equations. Syst Control Lett 104:78–85

    Article  MathSciNet  Google Scholar 

  3. Araruna FD, Fernández-Cara E, Santos MC (2015) Stackelberg–Nash exact controllability for linear and semilinear parabolic equations. ESAIM Control Optim Calc Var 21(3):835–856

    Article  MathSciNet  Google Scholar 

  4. Black F, Scholes M (1973) The pricing of options and corporate liabilities. J Political Econ 81:637–659

    Article  MathSciNet  Google Scholar 

  5. Cannarsa P, Martinez P, Vancostenoble J (2008) Carleman estimates for a class of degenerate parabolic equations. SIAM J Control Optim 47(1):1–19

    Article  MathSciNet  Google Scholar 

  6. Cannarsa P, Martinez P, Vancostenoble J (2016) Global Carleman estimates for degenerate parabolic operators with applications, vol 239. American Mathematical Society, Providence, p 1133

    MATH  Google Scholar 

  7. Cannarsa P, Martinez P, Vancostenoble J (2005) Null controllability of degenerate heat equations. Adv Differ Equ 10(2):153–190

    MathSciNet  MATH  Google Scholar 

  8. Cannarsa P, de Teresa L (2009) Controllability of 1-D coupled degenerate parabolic equations. Electron J Differ Equ 2009(73):1–21

  9. Díaz JI, Lions JL (2004) On the approximate controllability of Stackelberg–Nash strategies. In: Díaz JI (ed) Ocean circulation and pollution control – a mathematical and numerical investigation. Springer, Berlin, pp 17–27

    Chapter  Google Scholar 

  10. Doubova A, Fernández-Cara E, González-Burgos M, Zuazua E (2002) On the controllability of parabolic systems with a nonlinear term involving the state and the gradient. SIAM J Control Optim 41(3):798–819

    Article  MathSciNet  Google Scholar 

  11. Escauriaza L, Seregin G, Severak V (2003) Backward uniqueness for parabolic equations. Arch Ration Mech Anal 169:147–157

    Article  MathSciNet  Google Scholar 

  12. Escauriaza L, Seregin G, Severak V (2004) Backward uniqueness for the heat operator in half-space. St. Petersburg Math J 15:139–148

    Article  MathSciNet  Google Scholar 

  13. Fernández-Cara E, Guerrero S, Imanuvilov OY, Puel J-P (2004) Local exact controllability of the Navier–Stokes system. J Math Pures Appl 83(12):1501–1542

    Article  MathSciNet  Google Scholar 

  14. Flores C, de Teresa L (2010) Carleman estimates for degenerate parabolic equations with first order terms and applications. C. R. Acad Sci Paris Ser. I 348:391–396

    Article  MathSciNet  Google Scholar 

  15. Fursilkov AV, Ymanuvilov OY (1996) Controllability of evolution equations, Lecture Note Ser, vol 34. Seoul National University, Korea, Research Institute of Mathematics

  16. González-Burgos M, de Teresa L (2007) Some results on controllability for linear and nonlinear heat equations in unbounded domains. Adv Differ Equ 12(11):1201–1240

    MathSciNet  MATH  Google Scholar 

  17. Guillén-González F, Marques-Lopes FP, Rojas-Medar MA (2013) On the approximate controllability of Stackelberg–Nash strategies for Stokes equations. Proc Am Math Soc 141:1759–1773

    Article  MathSciNet  Google Scholar 

  18. Hernández-Santamaría V, de Tereza L, Poznyak A (2016) Hierarchic control for a coupled parabolic system. Portugal Math 73(2):115–137

    Article  MathSciNet  Google Scholar 

  19. Imanuvilov OY (2001) Remarks on exact controllability for the Navier–Stokes equations. ESAIM Control Optim Calc Var 6:39–72

    Article  MathSciNet  Google Scholar 

  20. Imanuvilov OY, Yamamoto M (2003) Carleman inequalities for parabolic equations in Sobolev spaces of negative order and exact controllability for semilinear parabolic equations. Publ Res Inst Math Sci 39(2):227–274

    Article  MathSciNet  Google Scholar 

  21. Mercan M (2013) Optimal control for distributed linear systems subjected to null-controllability. Appl Anal 92(9):1928–1943

    Article  MathSciNet  Google Scholar 

  22. Micu S, Zuazua E (2001) On the lack of null controllability of the heat equation on the half-line. Trans Am Math Soc 353:1635–1659

    Article  MathSciNet  Google Scholar 

  23. Nakoulima O (2007) Optimal control for distributed systems subject to null controllability. Application to discriminating sentinels. ESAIM Control Optim Calc Var 13(4):623–638

    Article  MathSciNet  Google Scholar 

  24. Wang C, Du R (2014) Carleman estimates and null controllability for a class of degenerate parabolic equations with convection terms. SIAM J Control Optim 52(3):1457–1480

    Article  MathSciNet  Google Scholar 

Download references

Acknowledgements

Most of this research was done while the first and second authors were visiting IMUS, at Universidad de Sevilla (Spain) and, then, while the third author was visiting DM, at Universidade Federal da Paraíba (João Pessoa - PB, Brazil). The authors express to these institutions their thanks for their assistance and hospitality.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to F. D. Araruna.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This work has been partially supported by INCTMat, CAPES, CNPq (Brazil), MathAmSud COSIP and Grant MTM2016-76991-P (MINECO, Spain).

Appendices

A Appendix: Proof of Theorem 1

It is sufficient to prove Theorem 1 assuming that \(b_0=0\).

Let v be the solution to (6) (where \(v_T\in L^2(0,1)\) and \(g\in L^2(Q)\)). For any \(s\ge s_0>0\), we set \(z=e^{-s\sigma }v\). By a density argument, we can assume without loss of generality that v is regular enough. We have

$$\begin{aligned} v_t=e^{s\sigma }(s\sigma _tz+z_t), \ \ (x^\alpha v_x)_x=e^{s\sigma }[s^2\sigma _x^2x^\alpha z+2s\sigma _xx^\alpha z_x+s(\sigma _xx^\alpha )_xz+(x^\alpha z_x)_x] \end{aligned}$$

and, consequently,

$$\begin{aligned} P^+z+P^-z=e^{-s\sigma }g, \end{aligned}$$

where \(P^+z=s\sigma _tz+s^2x^\alpha \sigma _x^2z+(x^\alpha z_x)_x\) and \(P^-z=z_t+s(x^\alpha \sigma _x)_xz+2sx^\alpha \sigma _xz_x.\) This gives

$$\begin{aligned} \Vert e^{-s\sigma }g\Vert ^2=\Vert P^+z\Vert ^2+\Vert P^-z\Vert ^2+2(\!( P^+z,P^-z)\!). \end{aligned}$$
(42)

We have that \((\!( P^+z,P^-z)\!)=I_1+\cdots +I_4\), where

$$\begin{aligned} I_1= & {} \left( \!\left( s\sigma _tz+s^2\sigma _x^2x^\alpha z+\left( x^\alpha z_x\right) _x,z_t\right) \!\right) \\ I_2= & {} s^2\left( \!\left( \sigma _tz,\left( x^\alpha \sigma _x\right) _xz+2\sigma _x x^\alpha z_x\right) \!\right) \\ I_3= & {} s^3\left( \!\left( \sigma _x^2x^\alpha z, \left( x^\alpha \sigma _x\right) _xz+2\sigma _x x^\alpha z_x\right) \!\right) \\ I_4= & {} s\left( \!\left( \, \left( x^\alpha z_x\right) _x, \left( x^\alpha \sigma _x\right) _xz+2\sigma _x x^\alpha z_x\right) \!\right) . \end{aligned}$$

The next step is to compute \(I_1,\ I_2, \ I_3 \) and \(I_4\). To this purpose, we will use that \(z=z_x=0\) at \(t=0\) and \(t=T\) and, also,

$$\begin{aligned} \displaystyle \int \!\!\!\!\int _Q(x^\alpha z_x)_xz_t\,\mathrm{d}x\,\mathrm{d}t=0. \end{aligned}$$

After integrating by parts, we deduce easily that

$$\begin{aligned} I_1= & {} -\frac{s}{2}\int \!\!\!\!\int _Q|z|^2(\sigma _{tt}+2s\sigma _x\sigma _{xt}x^\alpha )\,\mathrm{d}x\,\mathrm{d}t\\ I_2= & {} -s^2\int \!\!\!\!\int _Qx^\alpha \sigma _x\sigma _{xt}|z|^2\mathrm{d}x\,\mathrm{d}t;\\ I_3= & {} -s^3\int \!\!\!\!\int _Qx^\alpha \sigma _x(x^\alpha \sigma _x^2)_x|z|^2\,\mathrm{d}x\,\mathrm{d}t;\\ I_4= & {} -s\int \!\!\!\!\int _Q\left( x^\alpha \sigma _x\right) _{xx}x^\alpha zz_x\mathrm{d}x\,\mathrm{d}t-2s\int \!\!\!\!\int _Q\left( x^\alpha \sigma _x\right) _xx^\alpha |z_x|^2\,\mathrm{d}x\,\mathrm{d}t\\&+\,s\alpha \int \!\!\!\!\int _Q\sigma _xx^{2\alpha -1}|z_x|^2\,\mathrm{d}x\,\mathrm{d}t +\,\left. s\int _0^T\sigma _xx^{2\alpha }|z_x|^2\mathrm{d}t\right| _{x=0}^{x=1}. \end{aligned}$$

Consequently,

$$\begin{aligned} (\!( P^+z,P^-z)\!)\!= & {} \!\!\left. s\!\int _0^T\!\!\sigma _xx^{2\alpha }|z_x|^2\,\mathrm{d}t\right| _{x=0}^{x=1}\!-s^3\!\!\int \!\!\!\!\int _Q\!x^\alpha \sigma _x(x^\alpha \sigma _x^2)_x|z|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad -2s\!\int \!\!\!\!\int _Q\!(x^\alpha \sigma _x)_xx^\alpha |z_x|^2\,\mathrm{d}x\,\mathrm{d}t -2s^2\int \!\!\!\!\int _Q\!\!\sigma _x\sigma _{xt}x^\alpha |z|^2\,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad +\alpha s\int \!\!\!\!\int _Q\!\sigma _xx^{2\alpha -1}|z_x|^2\,\mathrm{d}x\,\mathrm{d}t-s\int \!\!\!\!\int _Q\!\!(x^\alpha \sigma _x)_{xx}x^\alpha zz_x\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad -\frac{s}{2}\int \!\!\!\!\int _Q\!\!\sigma _{tt}|z|^2\,\mathrm{d}x\,\mathrm{d}t. \end{aligned}$$
(43)

Using the definitions of the functions \(\sigma \) and \(\xi \), we see that

$$\begin{aligned} \sigma _x\sigma _{xt}x^\alpha= & {} \lambda ^2x^\alpha |\eta '|^2\xi \xi _t\\ x^\alpha \sigma _x(x^\alpha \sigma _x^2)_x= & {} -\,\lambda ^3x^\alpha \eta '(x^\alpha |\eta '|^2)_x\xi ^3-\,2\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3\\ (x^\alpha \sigma _x)_xx^\alpha= & {} -\,\lambda x^\alpha (x^\alpha \eta ')_x\xi -\,\lambda ^2x^{2\alpha }|\eta '|^2\xi \\ \sigma _xx^{2\alpha -\,1}= & {} -\,\lambda x^{2\alpha -\,1}\eta '\xi \\ (x^\alpha \sigma _x)_{xx}x^\alpha= & {} -\,\lambda x^\alpha (x^\alpha \eta ')_{xx}\xi -\,2\lambda ^2x^\alpha \eta '(x^\alpha \eta ')_x\xi -\,\lambda ^2x^{2\alpha }\eta '\eta ''\xi -\,\lambda ^3x^{2\alpha }(\eta ')^3\xi \\ \sigma _xx^{2\alpha }= & {} -\,\lambda x^{2\alpha }\eta '\xi . \end{aligned}$$

With this information, we will now estimate each term in the right-hand side of (43). For \(s_0\) and \(\lambda _0\) large enough, we have

$$\begin{aligned} \left. s\int _0^T\!\!\sigma _xx^{2\alpha }|z_x|^2\,\mathrm{d}t\right| _{x=0}^{x=1}\ge & {} 0 , \end{aligned}$$
(44)
$$\begin{aligned} -\,s^3\int \!\!\!\!\int _Q\!\!x^\alpha \sigma _x(x^\alpha \sigma _x^2)_x|z|^2\,\mathrm{d}x\,\mathrm{d}t\ge & {} C\left( s^3\lambda ^4\int \!\!\!\!\int _Q\!\!x^{2\alpha }|\eta '|^4\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t\right. \nonumber \\&\left. +\,s^3\lambda ^3\int \!\!\!\!\int _{Q_0}\!\!\!x^{2-\,\alpha }\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t\right) \nonumber \\&\quad \!\!\!\!\!-\,Cs^3\lambda ^3\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t, \end{aligned}$$
(45)
$$\begin{aligned} -\,2s\int \!\!\!\!\int _Qx^\alpha (x^\alpha \sigma _x)_x|z_x|^2\,\mathrm{d}x\,\mathrm{d}t\ge & {} Cs\lambda ^2\int \!\!\!\!\int _Qx^{2\alpha }|\eta '|^2\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\!\!\!+\,2s\lambda \int \!\!\!\!\int _{Q_0}\!\!\!x^\alpha \xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t-\,Cs\lambda \!\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t, \nonumber \\ \end{aligned}$$
(46)
$$\begin{aligned} -\,2s^2\int \!\!\!\!\int _Qx^\alpha \sigma _x\sigma _{xt}|z|^2\,\mathrm{d}x\,\mathrm{d}t\!\!\ge & {} \!\! -\,Cs^2\lambda ^2\left( \int \!\!\!\!\int _Qx^{2\alpha }|\eta '|^4\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t \right. \nonumber \\&\!\left. +\int \!\!\!\!\int _{Q_0}\!\!\!x^{2-\,\alpha }\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t +\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t\right) \nonumber \\ \end{aligned}$$
(47)

and

$$\begin{aligned} \alpha s\int \!\!\!\!\int _Qx^{2\alpha -\,1}\sigma _x|z_x|^2\,\mathrm{d}x\,\mathrm{d}t \ge -\,\alpha s\lambda \!\int \!\!\!\!\int _{Q_0}\!\!\!x^\alpha \xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t-\,C s\lambda \!\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\!\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t,\nonumber \\ \end{aligned}$$
(48)

where C only depends on \(a_0\), T and \(\alpha \). On the other hand,

$$\begin{aligned} -\,s\!\int \int _Q\!x^\alpha (x^\alpha \sigma _x)_{xx}zz_x\mathrm{d}x\,\mathrm{d}t= & {} s\lambda \!\int \int _Q\!x^\alpha (x^\alpha \eta ')_{xx}\xi zz_x\mathrm{d}x\,\mathrm{d}t\nonumber \\&+s\lambda ^2\!\int \!\!\!\!\int _Q\!x^{2\alpha }\eta '\eta ''\xi zz_x\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad +\,2s\lambda ^2\!\int \!\!\!\!\int _Q\!x^\alpha \eta '(x^\alpha \eta ')_x\xi zz_x\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad +\,s\lambda ^3\!\int \!\!\!\!\int _Q\!x^{2\alpha }(\eta ')^3\xi zz_x\mathrm{d}x\,\mathrm{d}t . \end{aligned}$$
(49)

In order to estimate this last term, we note that

$$\begin{aligned} s\lambda \int \!\!\!\!\int _Qx^\alpha (x^\alpha \eta ')_{xx}\xi zz_x\,\mathrm{d}x\,\mathrm{d}t\ge & {} -\,Cs\lambda \int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\left( \xi ^3|z|^2+\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t, \end{aligned}$$
(50)
$$\begin{aligned} s\lambda ^2\int \!\!\!\!\int _Qx^{2\alpha }\eta '\eta ''\xi zz_x\,\mathrm{d}x\,\mathrm{d}t\ge & {} -\,C\int \!\!\!\!\int _Q\left( s^2\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|z|^2+x^{2\alpha }|\eta '|^2\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\nonumber \\&-\,C\int \!\!\!\!\int _{Q_0}\!\!\!\left( s^2\lambda ^3x^{2-\,\alpha }\xi ^3|z|^2+\lambda x^\alpha \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\nonumber \\&-\,C\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\left( s^2\lambda ^3\xi ^3 |z|^2+\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t, \end{aligned}$$
(51)
$$\begin{aligned} 2s\lambda ^2\int \!\!\!\!\int _Qx^\alpha \eta '(x^\alpha \eta ')_x\xi zz_x\,\mathrm{d}x\,\mathrm{d}t\ge & {} -\,C\int \!\!\!\!\int _Q\left( s^2\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|z|^2+x^{2\alpha }|\eta '|^2\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\nonumber \\&-\,C\int \!\!\!\!\int _{Q_0}\!\!\!\left( s^2\lambda ^3x^{2-\,\alpha }\xi ^3|z|^2+\lambda x^\alpha \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\nonumber \\&-\,C\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\left( s^2\lambda ^3\xi ^3 |z|^2+\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \end{aligned}$$
(52)

and

$$\begin{aligned} s\lambda ^3\!\int \!\!\!\!\int _Q\!x^{2\alpha }(\eta ')^3\xi zz_x\,\mathrm{d}x\,\mathrm{d}t\!\!\ge & {} \!\!-\,C\int \!\!\!\!\int _Q\left( \lambda ^2x^{2\alpha }|\eta '|^2\xi |z_x|^2+s^2\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3 |z|^2\right) \,\mathrm{d}x\,\mathrm{d}t.\nonumber \\ \end{aligned}$$
(53)

By (49)–(53), we conclude that

$$\begin{aligned}&-\,s\int \!\!\!\!\int _Qx^\alpha (x^\alpha \sigma _x)_{xx}zz_x\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad \ge -\,C\int \!\!\!\!\int _Q\left( s^2\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|z|^2+\lambda ^2x^{2\alpha }|\eta '|^2\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\qquad -\,C\int \!\!\!\!\int _{Q_0}\!\!\!\left( s^2\lambda ^3x^{2-\,\alpha }\xi ^3|z|^2+\lambda x^\alpha \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\qquad -\,C\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\left( s^2\lambda ^3\xi ^3|z|^2+s\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t. \end{aligned}$$
(54)

From (43)–(48) and (54), we conclude that

$$\begin{aligned} (\!( P^+z,P^-z)\!)\ge & {} C\left( \int \!\!\!\!\int _Q\left( s^3\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|z|^2+s\lambda ^2x^{2\alpha }|\eta '|^2\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \right. \\&\left. +\int _{Q_0}\!\!\left( s^3\lambda ^3x^{2-\,\alpha }\xi ^3|z|^2+s\lambda x^\alpha \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\right) \nonumber \\&-\,C\left( \int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\left( s^3\lambda ^3\xi ^3|z|^2+s\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t+s\int \!\!\!\!\int _Q\xi ^{3/2}|z|^2\,\mathrm{d}x\,\mathrm{d}t\right) . \end{aligned}$$

Using (42), we find that

$$\begin{aligned}&\Vert P^+z\Vert ^2+\Vert P^-\,z\Vert ^2+\int \!\!\!\!\int _Q\left( s^3\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|z|^2+s\lambda ^2x^{2\alpha }|\eta '|^2\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad +\int \!\!\!\!\int _{Q_0}\!\!\left( s^3\lambda ^3x^{2-\,\alpha }\xi ^3|z|^2+s\lambda x^\alpha \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad \le C\left( \Vert e^{-\,s\sigma }g\Vert ^2+\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\left( s^3\lambda ^3\xi ^3|z|^2+s\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t+s\int \!\!\!\!\int _Q\xi ^{3/2}|z|^2\,\mathrm{d}x\,\mathrm{d}t\right) .\nonumber \\ \end{aligned}$$
(55)

Furthermore, we see that

$$\begin{aligned} s^3\lambda ^3\int \!\!\!\!\int _Q\xi ^3x^{2-\,\alpha }|z|^2\,\mathrm{d}x\,\mathrm{d}t\le & {} Cs^3\lambda ^3\left( \int \!\!\!\!\int _{Q_0\!}\xi ^3x^{2-\,\alpha }|z|^2\,\mathrm{d}x\,\mathrm{d}t\right. \\&\left. +\int \!\!\!\!\int _Q\xi ^3x^{2\alpha }|\eta '|^4|z|^2\,\mathrm{d}x\,\mathrm{d}t+\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t\right) \end{aligned}$$

and

$$\begin{aligned} s\lambda \int \!\!\!\!\int _Q\xi x^\alpha |z_x|^2\,\mathrm{d}x\,\mathrm{d}t\le & {} Cs\lambda \left( \int \!\!\!\!\int _{Q_0}\!\xi x^\alpha |z_x|^2\,\mathrm{d}x\,\mathrm{d}t+\int \!\!\!\!\int _Q\xi x^{2\alpha }|\eta '|^2|z_x|^2\,\mathrm{d}x\,\mathrm{d}t\right. \\&\left. +\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t\right) . \end{aligned}$$

Hence, from (55), we obtain:

$$\begin{aligned}&\Vert P^+z\Vert ^2+\Vert P^-\,z\Vert ^2+\int \!\!\!\!\int _Q\left( s^3\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|z|^2+s\lambda ^2x^{2\alpha }|\eta '|^2\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\qquad +\int \!\!\!\!\int _Q\!\!\left( s^3\lambda ^3x^{2-\,\alpha }\xi ^3|z|^2+s\lambda x^\alpha \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad \le C\left( \Vert e^{-\,s\sigma }g\Vert ^2+\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\left( s^3\lambda ^3\xi ^3|z|^2+s\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t+s\int \!\!\!\!\int _Q\xi ^{3/2}|z|^2\,\mathrm{d}x\,\mathrm{d}t\right) .\nonumber \\ \end{aligned}$$
(56)

Let us denote by L(z) all the terms in the left-hand side of (56). For instance, assume that \(\alpha \ne 1\). Using Young and Hardy’s Inequality we deduce that

$$\begin{aligned} s^2\lambda ^2\int \!\!\!\!\int _Q\xi ^2|z|^2\,\mathrm{d}x\,\mathrm{d}t\le & {} s^3\lambda ^3\int \!\!\!\!\int _Q\xi ^3x^{2-\alpha }|z|^2\,\mathrm{d}x\,\mathrm{d}t+s\lambda \int \!\!\!\!\int _Qx^{\alpha -2}|\xi ^{1/2}z|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\\le & {} s^3\lambda ^3\int \!\!\!\!\int _Q\xi ^3x^{2-\alpha }|z|^2\,\mathrm{d}x\,\mathrm{d}t+s\lambda \int \!\!\!\!\int _Qx^\alpha |(\xi ^{1/2}z)_x|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\\le & {} CL(z). \end{aligned}$$
(57)

Now, let us assume that \(\alpha =1\). Using Hölder, Young and Hardy inequalities, we see that

$$\begin{aligned} s^{3/2}\lambda ^{7/4}\int \!\!\!\!\int _Q\xi ^{3/2}|z|^2\,\mathrm{d}x\,\mathrm{d}t= & {} \int \!\!\!\!\int _{Q_0}(s^3\lambda ^4x^2\xi ^3|z|^2)^{1/4}.(s\lambda x^{-2/3}\xi |z|^2)^{3/4}\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&+s^{3/2}\lambda ^{7/4}\int _0^T\!\!\!\int _{a_0}^1\xi ^{3/2}|z|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\\le & {} \quad Cs^3\lambda ^4\int \!\!\!\!\int _Q\xi ^3x^{2\alpha }|\eta '|^4|z|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&+s\lambda \int \!\!\!\!\int _Qx^{-\,2+4/3}(\xi ^{1/2}z)^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\\le & {} CL(z). \end{aligned}$$
(58)

From (56), (57) and (58), we deduce that

$$\begin{aligned}&\Vert P^+z\Vert ^2+\Vert P^-z\Vert ^2+\int \!\!\!\!\int _Q\left( s^3\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|z|^2+s\lambda ^2x^{2\alpha }|\eta '|^2\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\qquad +\int \!\!\!\!\int _Q\!\!\left( s^3\lambda ^3x^{2-\alpha }\xi ^3|z|^2+s^2\lambda ^2\gamma _1(\lambda )\xi ^2\gamma _2(s\xi )|z|^2+s\lambda x^\alpha \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad \le C\left( \Vert e^{-s\sigma }g\Vert ^2+\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\left( s^3\lambda ^3\xi ^3|z|^2+s\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\right) . \end{aligned}$$
(59)

Now, we will work to include the terms with a first-order time derivative and second-order spatial derivatives in the left-hand side. Using the estimate (59) and the definitions of \(P^-z\) and \(P^+z\), we have

$$\begin{aligned}&s^{-1}\gamma _1(\lambda )\int \!\!\!\!\int _Q\xi ^{-1}|z_t|^2\,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad \le C\left( \Vert e^{-s\sigma }g\Vert ^2+\int \!\!\!\!\int _{\omega _{0T}}\left( s^3\lambda ^3\xi ^3|z|^2+s\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\right) \end{aligned}$$
(60)

and

$$\begin{aligned}&s^{-1}\gamma _1(\lambda )\!\!\int \!\!\!\!\int _Q\xi ^{-1}|(x^\alpha z_x)_x|^2\,\mathrm{d}x\,\mathrm{d}t \le C\left( \Vert e^{-s\sigma }g\Vert ^2+\int \!\!\!\!\int _{\omega _{0T}}\left( s^3\lambda ^3\xi ^3|z|^2+s\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\right) \!.\nonumber \\ \end{aligned}$$
(61)

By (59), (60) and (61), it is now clear that

$$\begin{aligned}&\int \!\!\!\!\int _Q\left[ s^{-1}\gamma (\lambda )\xi ^{-1}\left( |z_t|^2+|(x^\alpha z_x)_x|^2\right) +s\lambda x^\alpha \xi |z_x|^2+s\lambda ^2x^{2\alpha }|\eta '|^2\xi |z_x|^2\right] \,\mathrm{d}x\,\mathrm{d}t \\&\qquad +\int \!\!\!\!\int _Q\left( s^2\lambda ^2\gamma _1(\lambda )\xi ^2\gamma _2(s\xi )|z|^2+s^3\lambda ^3x^{2-\alpha }\xi ^3|z|^2+s^3\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|z|^2\right) \,\mathrm{d}x\,\mathrm{d}t\\&\quad \le C\left( \Vert e^{-s\sigma }g\Vert ^2+\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\left( s^3\lambda ^3\xi ^3|z|^2+s\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\right) . \end{aligned}$$

Coming back to the original variable v and using the estimate

$$\begin{aligned} \int _{\omega _{0T}}\!\!\!\!e^{-2s\sigma }\xi |v_x|^2\mathrm{d}x\,\mathrm{d}t\le & {} Cs^2\!\lambda ^2\int \!\!\!\!\int _{\omega _T}\!\!\!\!e^{-2s\sigma }\xi ^3 |v|^2\mathrm{d}x\,\mathrm{d}t\nonumber \\&+\,Cs^{-2}\lambda ^{-2}\int \!\!\!\!\int _Qe^{-2s\sigma }\xi ^{-1}|(x^\alpha v_x)_x|^2\,\mathrm{d}x\,\mathrm{d}t, \end{aligned}$$

we find the desired inequality (20).\(\Box \)

B Appendix: Proof of Theorem 2

Again, we will assume that \(b_0=0\). Let us set \(g_0:=g-b_1v_x\). Arguing as before, we can deduce that

$$\begin{aligned} (\!( P^+z,P^-z)\!)= & {} s\!\left. \int _0^T\!\sigma _xx^{2\alpha }|z_x|^2\,\mathrm{d}t\right| _{x=0}^{x=1}-s^3\!\int \!\!\!\!\int _Qx^\alpha \sigma _x(x^\alpha \sigma _x^2)_x|z|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&-2s\!\int \!\!\!\!\int _Q\!(x^\alpha \sigma _x)_xx^\alpha |z_x|^2\,\mathrm{d}x\,\mathrm{d}t-2s^2\!\int \!\!\!\!\int _Q\!\sigma _x\sigma _{xt}x^\alpha |z|^2\,\mathrm{d}x\,\mathrm{d}t \nonumber \\&+\alpha s\!\int \!\!\!\!\int _Q\!\sigma _xx^{2\alpha -1}|z_x|^2\,\mathrm{d}x\,\mathrm{d}t-\frac{s}{2}\int \!\!\!\!\int _Q\sigma _{tt}|z|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&-s\int \!\!\!\!\int _Q(x^\alpha \sigma _x)_{xx}x^\alpha zz_x\,\mathrm{d}x\,\mathrm{d}t. \end{aligned}$$
(62)

Moreover, estimating the terms on the right-hand side of (62), we obtain that

$$\begin{aligned} (\!( P^+z,P^-z)\!)\ge & {} C\left[ s^3\lambda ^4\int \!\!\!\!\int _Qx^{2\alpha }|\eta '|^4\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t+s^3\lambda ^3\int \!\!\!\!\int _{Q_0}\!\!\!\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t\right. \nonumber \\&\left. s\lambda ^2\int \!\!\!\!\int _Qx^{2\alpha }|\eta '|^2\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t\right] \nonumber \\&+\frac{1}{3}(2-\alpha )s\lambda \int \!\!\!\!\int _{Q_0}\!\!x^{(4\alpha -2)/3}\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t \nonumber \\&-C\left[ s^3\lambda ^3\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t+s\lambda \int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t\right] \nonumber \\&-s\int \!\!\!\!\int _Q(x^\alpha \sigma _x)_{xx}x^\alpha zz_x\,\mathrm{d}x\,\mathrm{d}t. \end{aligned}$$
(63)

Now, we can estimate the last term in the right-hand side of (63). It is just here where the restriction \(\alpha <1/2\) is required. We have

$$\begin{aligned} -s\!\int \!\!\!\!\int _Qx^\alpha (x^\alpha \sigma _x)_{xx}zz_x\,\mathrm{d}x\,\mathrm{d}t= & {} s\lambda ^2\!\!\int \!\!\!\!\int _Qx^{2\alpha }\eta '\eta ''\xi zz_x\,\mathrm{d}x\,\mathrm{d}t+s\lambda ^3\!\!\int \!\!\!\!\int _Qx^{2\alpha }(\eta ')^3\xi zz_x\,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad +2s\lambda ^2\!\!\int \!\!\!\!\int _Qx^\alpha \eta '(x^\alpha \eta ')_x\xi zz_x\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad +s\lambda \!\int \!\!\!\!\int _Qx^\alpha (x^\alpha \eta ')_{xx}\xi zz_x\,\mathrm{d}x\,\mathrm{d}t \nonumber \\\ge & {} -C\int \!\!\!\!\int _Q\left( s^2\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|z|^2+x^{2\alpha }|\eta '|^2\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad -C\int \!\!\!\!\int _{Q_0}\!\!\!\left( s^2\lambda ^3\xi ^3|z|^2+\lambda x^{(4\alpha -2)/3}\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad -C\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\left( s^2\lambda ^3\xi ^3 |z|^2+\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad + s\lambda \!\int \!\!\!\!\int _Qx^\alpha (x^\alpha \eta ')_{xx}\xi zz_x\,\mathrm{d}x\,\mathrm{d}t. \end{aligned}$$
(64)

In the previous case we had \((x^\alpha \eta ')_{xx}=0\) in \((0,1)\backslash \omega _0\). Here, this is lost and, therefore, in order to estimate the last term in the right-hand side of (64), we have to work differently.

Using integration by parts, we have that

$$\begin{aligned}&s\lambda \!\int \!\!\!\!\int _Q\!x^\alpha (x^\alpha \eta ')_{xx}\xi zz_x\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad =-\int \!\!\!\!\int _Q\frac{s\lambda }{2}\xi |z|^2\left[ \lambda x^\alpha \eta '(x^\alpha \eta ')_{xx}+[x^\alpha (x^\alpha \eta ')_{xx}]_x\right] \,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad \ge Cs\lambda ^2\!\int \!\!\!\!\int _{Q_0}\!\!\!x^{(2\alpha -4)/3}\xi |z|^2\,\mathrm{d}x\,\mathrm{d}t-Cs\lambda ^2\!\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\xi |z|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\qquad -\,Cs\lambda ^2\!\int \!\!\!\!\int _{(b_0,1)\times (0,T)}x^{(2\alpha -4)/3}\xi |z|^2\,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\qquad -\frac{s\lambda }{2}\int \!\!\!\!\int _Q\xi |z|^2[x^\alpha (x^\alpha \eta ')_{xx}]_x\,\mathrm{d}x\,\mathrm{d}t. \end{aligned}$$
(65)

In view of Proposition 2, we have

$$\begin{aligned}&-\frac{s\lambda }{2}\!\int \!\!\!\!\int _Q[x^\alpha (x^\alpha \eta ')_{xx}]_x\xi |z|^2\,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad \ge \ -\frac{2}{3}s\lambda \frac{(1+\alpha )(2-\alpha )}{5-4\alpha }\int \!\!\!\!\int _{Q_0}\!\!x^{(4\alpha -2)/3}(\xi ^{1/2}z)_x^2\,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\qquad -\,Cs\lambda \int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\xi ^3 |z|^2\,\mathrm{d}x\,\mathrm{d}t-\,Cs\lambda \int \!\!\!\!\int _Qx^{2\alpha }|\eta '|^4\xi ^3 |z|^2\,\mathrm{d}x\,\mathrm{d}t. \end{aligned}$$
(66)

Furthermore,

$$\begin{aligned}&\frac{2}{3}s\lambda \frac{(1+\alpha )(2-\alpha )}{5-4\alpha }\int \!\!\!\!\int _{Q_{0}}x^{(4\alpha -2)/3}(\xi ^{1/2}z)_{x}^{2}\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad =-m(\alpha )s\lambda \int \!\!\!\!\int _{Q_{0}}x^{(4\alpha -2)/3}\left[ \xi |z_{x}|^{2}+\lambda x^{(1-2\alpha )/3}\xi zz_{x}\!+\!\frac{\lambda ^{2}}{4}x^{(2-4\alpha )/3}\xi |z|^{2}\right] \,\mathrm{d}x\,\mathrm{d}t,\nonumber \\ \end{aligned}$$
(67)

where \(m(\alpha ):=\frac{2}{3}(1+\alpha )(2-\alpha )(5-4\alpha )^{-1}\) and

$$\begin{aligned} -Cs\lambda ^2\!\!\int \!\!\!\!\int _{Q_0}\!\!\!x^{(2\alpha -1)/3}\xi zz_x\,\mathrm{d}x\,\mathrm{d}t\ge & {} -C\lambda \!\int \!\!\!\!\int _{Q_0}\!\!\!\left( \lambda x^{(4\alpha -2)/3}\xi |z_x|^2+s^2\lambda ^3\xi ^3|z|^2\right) \,\mathrm{d}x\,\mathrm{d}t.\nonumber \\ \end{aligned}$$
(68)

Then from (66), (67) and (68), we get

$$\begin{aligned} -\frac{s\lambda }{2}\int \!\!\!\!\int _Q[x^\alpha (x^\alpha \eta ')_{xx}]_x\xi |z|^2\,\mathrm{d}x\,\mathrm{d}t\ge & {} \ -m(\alpha )s\lambda \int \!\!\!\!\int _{Q_0}\!\!\!x^{(4\alpha -2)/3}\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&-Cs^2\lambda ^3\!\!\int \!\!\!\!\int _{Q_0}\!\!\!\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&-Cs\lambda \!\int \!\!\!\!\int _Qx^{2\alpha }|\eta '|^4\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&-C\lambda \!\int \!\!\!\!\int _{Q_0}\!\!\!x^{(4\alpha -2)/3}\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&-Cs\lambda \!\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t. \end{aligned}$$
(69)

Hence, from (69) and (65) we deduce that

$$\begin{aligned} s\lambda \!\int \!\!\!\!\int _Qx^\alpha (x^\alpha \eta ')_{xx}\xi zz_x\,\mathrm{d}x\,\mathrm{d}t\ge & {} Cs\lambda ^2\!\!\int \!\!\!\!\int _{Q_0}\!\!\!x^{(2\alpha -4)/3}\xi |z|^2\,\mathrm{d}x\,\mathrm{d}t-Cs^2\lambda ^3\!\!\int \!\!\!\!\int _{Q_0}\!\!\!\xi ^3|z|^2\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad -m(\alpha )s\lambda \!\int \!\!\!\!\int _{Q_0}\!\!\!x^{(4\alpha -2)/3}\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad -C\lambda \!\int \!\!\!\!\int _{Q_0}\!\!\!x^{(4\alpha -2)/3}\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad -Cs\lambda ^2\!\!\int \!\!\!\!\int _Qx^{2\alpha }|\eta '|^4\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad -Cs\lambda ^2\!\!\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\xi ^3|z|^2\,\mathrm{d}x\,\mathrm{d}t. \end{aligned}$$
(70)

Note that the new strategy makes appear an additional term in the right-hand side of (70).

From (64) and (70), we conclude that

$$\begin{aligned} -s\!\int \!\!\!\!\int _Qx^\alpha (x^\alpha \sigma _x)_{xx}zz_x\,\mathrm{d}x\,\mathrm{d}t\ge & {} -C\int \!\!\!\!\int _Q\left( \lambda ^2x^{2\alpha }|\eta '|^2\xi |z_x|^2+s^2\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|z|^2\right) \,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad +Cs\lambda ^2\!\!\int \!\!\!\!\int _{Q_0}\!\!\!x^{(2\alpha -4)/3}\xi |z|^2\,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad -C\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\left( s^2\lambda ^3\xi ^3|z|^2+s\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad -C\int \!\!\!\!\int _{Q_0}\!\!\!\left( s^2\lambda ^3\xi ^3|z|^2+\lambda x^{(4\alpha -2)/3}\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\nonumber \\&\quad -m(\alpha )s\lambda \!\int \!\!\!\!\int _{Q_0}\!\!\!x^{(4\alpha -2)/3}\xi |z_x|^2\,\mathrm{d}x\,\mathrm{d}t. \end{aligned}$$
(71)

It is important to take care of the constants accompanying the term

$$\begin{aligned} s\lambda \int \!\!\!\!\int _{Q_0}\!\!\!x^{(4\alpha -2)/3}\xi |z_x|^2\mathrm{d}x\,\mathrm{d}t \end{aligned}$$

in the estimates (63) and (71). The sum of these constants is

$$\begin{aligned} \frac{(2-\alpha )(1-2\alpha )}{5-4\alpha }, \end{aligned}$$

that is only positive for \(\alpha \in [0,1/2)\).

From (62), (63) and (71), we deduce that

$$\begin{aligned}&\Vert P^+z\Vert ^2+\Vert P^-z\Vert ^2+\int \!\!\!\!\int _Q\left( s^3\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|z|^2+s\lambda ^2x^{2\alpha }|\eta '|^2\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\ \ \ \ \ \ \nonumber \\&\quad +\int \!\!\!\!\int _{Q_0}\!\!\!\left( s^3\lambda ^3\xi ^3|z|^2+s\lambda ^2x^{(2\alpha -4)/3}\xi |z|^2+s\lambda x^{(4\alpha -2)/3}\xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad \le C\left( \Vert e^{-s\sigma }g_0\Vert ^2+\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\left( s^3\lambda ^3\xi ^3|z|^2+s\lambda \xi |z_x|^2\right) \,\mathrm{d}x\,\mathrm{d}t\right) .\ \end{aligned}$$

Arguing as in Appendix A, we can replace the integral in \(Q_0\) in the left-hand side by integrals in Q. Moreover, thanks to the additional term in the left-hand side of this inequality, we can incorporate terms with time derivatives and second-order spatial derivatives more easily and also return to the variable v:

$$\begin{aligned}&\int \!\!\!\!\int _Qe^{-2s\sigma }\left[ s^{-1}\xi ^{-1}\left( |v_t|^2+|(x^\alpha v_x)_x|^2\right) +s\lambda x^{(4\alpha -2)/3}\xi |v_x|^2+s\lambda ^2x^{2\alpha }|\eta '|^2\xi |v_x|^2\right] \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\qquad +\int \!\!\!\!\int _Qe^{-2s\sigma }|v|^2\left[ s\lambda ^2x^{(2\alpha -4)/3}\xi +s^3\lambda ^3\xi ^3+s^3\lambda ^4x^{2\alpha }|\eta '|^4\xi ^3|v|^2\right] \,\mathrm{d}x\,\mathrm{d}t \nonumber \\&\quad \le C\left( \Vert e^{-s\sigma }(g-b_1v_x)\Vert ^2+\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!e^{-2s\sigma }\left[ s^3\lambda ^3\xi ^3|z|^2+s\lambda \xi |z_x|^2\right] \,\mathrm{d}x\,\mathrm{d}t\right) . \end{aligned}$$
(72)

Since the power of x in the local term with first-order spatial derivatives is negative, we deduce that (72) remains true with \((g-b_1v_x)\) replaced by g. Finally, to eliminate the term with derivatives in the right-hand side, it suffices to work as in the case considered in Appendix A. \(\Box \)

C Appendix: Proof of Theorem 3

Again we will prove Theorem 3 when \(B=0\). In view of the presence of Robin conditions in (12), we will perform another change of variables:

$$\begin{aligned} v(y,t)=e^{-\eta (y)}w(y,t). \end{aligned}$$

Now, (12) becomes

$$\begin{aligned} \left\{ \begin{array} [c]{lll} v_{t}+v_{yy}=g_{0} &{} \text{ in } &{} Q^{\prime },\\ \left( v_{x}-\frac{1}{2}v\right) (0,t)=0\ \ \text{ and }\ \ \displaystyle \lim _{y\rightarrow \infty }\left( v_{y}-\frac{1}{2}v\right) (y,t)=0 &{} \text{ on } &{} (0,T),\\ v(x,T)=v_{T}(x) &{} \text{ in } &{} (0,1), \end{array} \right. \end{aligned}$$
(73)

where \(g_0:=e^{-\eta }F-(\eta ''+|\eta '|^2)v-2\eta 'v_y\).

Let v be a solution to (73). For any \(s\ge s_0>0\), we set \(z=e^{-s\sigma }v\) and \(\tilde{z}=e^{-s\tilde{\sigma }}v\). We have that z and \(\tilde{z}\) satisfy the following initial, final and boundary conditions:

$$\begin{aligned}&z=\tilde{z}=z_y=\tilde{z}_y=0 \ \ \text{ at } \ \ t=0 \ \ \text{ and }\ \ t=T, \\&\left( z_y+\left( s\lambda \xi -\frac{1}{2}\right) z\right) (0,t)=0 \ \ \ \text{ and } \ \ \displaystyle \lim _{y\rightarrow +\,\infty }z_y(y,t)=\lim _{y\rightarrow +\,\infty }z(y,t)=0\ \ \text{ on } \ \ [0,T], \\&\left( \tilde{z}_y+\left( -s\lambda \tilde{\xi }-\frac{1}{2}\right) \tilde{z}\right) (0,t)=0\ \ \ \text{ and } \ \ \displaystyle \lim _{y\rightarrow +\,\infty }\tilde{z}_y(y,t)=\lim _{y\rightarrow +\,\infty }\tilde{z}(y,t)=0 \ \ \text{ on } \ \ [0,T]. \end{aligned}$$

Again, we assume that v is regular enough. We have:

$$\begin{aligned} v_t=e^{s\sigma }[s\sigma _tz+z_t], \ \ \ \ v_{yy}=e^{s\sigma }[z_{yy}+s^2\sigma _y^2z+2s\sigma _yz_y+s\sigma _{yy}z] \end{aligned}$$

and, consequently,

$$\begin{aligned} M_1z+M_2z=g_1, \end{aligned}$$

with \(M_1z:=-2s\lambda ^2|\eta '|^2\xi z-2s\lambda |\eta '|\xi z_y+z_t\), \(M_2z:=s^2\lambda ^2|\eta '|^2\xi ^2z+z_{yy}+s\sigma _t z\) and \(g_1:=e^{-s\sigma }g_0+s\lambda \eta ''\xi z-s\lambda ^2|\eta '|^2\xi z\). This gives

$$\begin{aligned} \Vert M_1z\Vert ^2+\Vert M_2z\Vert ^2+2(\!( M_1z,M_2z)\!)= & {} \Vert g_1\Vert ^2. \end{aligned}$$
(74)

After some work, we can deduce that

$$\begin{aligned} (\!( M_1z,M_2z)\!)\ge & {} C\int \!\!\!\!\int _{Q'}\left( s^3\lambda ^4\xi ^3|z|^2+s\lambda ^2\xi |z_x|^2\right) \,\mathrm{d}y\,\mathrm{d}t\nonumber \\&-C\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\!\left( s^3\lambda ^4\xi ^3|z|^2+s\lambda ^2\xi |z_y|^2\right) \,\mathrm{d}y\,\mathrm{d}t\nonumber \\&+\left. \int _0^T\left[ s^3\lambda ^3(\eta ')^3\xi ^3|z|^2-s\lambda ^3(\eta ')^3\xi |z|^2+s^2\lambda \eta '\xi \sigma _t|z|^2\right] \,\mathrm{d}t\right| _{y=0}\nonumber \\&+\left. \int _0^T\left( s\lambda \eta '\xi |z_y|^2+2s\lambda ^2|\eta '|^2\xi zz_y+zz_{yt}\right) \,\mathrm{d}t\right| _{y=0}. \end{aligned}$$
(75)

Working similarly with the function \(\tilde{z}\), we obtain

$$\begin{aligned} \tilde{M}_1\tilde{z}+\tilde{M}_2\tilde{z}=\tilde{g}_1 \end{aligned}$$

with \(\tilde{M}_1\tilde{z}:=\tilde{I}_{11}+\tilde{I}_{12}+\tilde{I}_{13}:=-2s\lambda ^2|\eta '|^2\tilde{\xi }\tilde{z}+2s\lambda \eta '\tilde{\xi }\tilde{z}_y+\tilde{z}_t\), \(\tilde{M}_2\tilde{z}:=\tilde{I}_{21}+\tilde{I}_{22}+\tilde{I}_{23}:=s^2\lambda ^2|\eta '|^2\tilde{\xi }^2\tilde{z}+\tilde{z}_{yy}+s\tilde{\sigma }_t\tilde{z}\) and \(\tilde{g}_1:=e^{-s\tilde{\sigma }}g_0-s\lambda \eta ''\tilde{\xi }\tilde{z}-s\lambda ^2|\eta '|^2\tilde{\xi }\tilde{z}\). This gives:

$$\begin{aligned} \Vert \tilde{g}_1\Vert ^2\!= & {} \!\Vert \tilde{M}_1\tilde{z}\Vert ^2+\Vert \tilde{M}_2\tilde{z}\Vert ^2+2(\!(\tilde{M}_1\tilde{z},\tilde{M}_2\tilde{z})\!) \end{aligned}$$
(76)

and

$$\begin{aligned} (\!(\tilde{M}_1\tilde{z},\tilde{M}_2\tilde{z})\!)\ge & {} C\int \!\!\!\!\int _{Q'}s^3\lambda ^4\tilde{\xi }^3|\tilde{z}|^2+s\lambda ^2\tilde{\xi }|\tilde{z}_y|^2\,\mathrm{d}y\,\mathrm{d}t\nonumber \\&-C\int \!\!\!\!\int _{\omega _{0T}}s^3\lambda ^4\tilde{\xi }^3|\tilde{z}|^2+s\lambda ^2\tilde{\xi }|\tilde{z}_y|^2\,\mathrm{d}y\,\mathrm{d}t\nonumber \\&+\left. \int _0^T[-s^3\lambda ^3(\eta ')^3\tilde{\xi }^3\tilde{z}^2+s\lambda ^3(\eta ')^3\tilde{\xi } \tilde{z}^2-s^2\lambda \eta '\tilde{\xi }\tilde{\sigma }_t\tilde{z}^2]\,\mathrm{d}t\right| _{y=0} \nonumber \\&+\left. \int _0^T\left( -s\lambda \eta '\tilde{\xi } \tilde{z}_y^2+2s\lambda ^2|\eta '|^2\tilde{\xi } \tilde{z}\tilde{z}_y+\tilde{z}\tilde{z}_{yt}\,\mathrm{d}t\right) \right| _{y=0}. \end{aligned}$$
(77)

Note that \(z=\tilde{z}\), \(\xi =\tilde{\xi }\), \(\sigma =\tilde{\sigma }\) and \(\sigma _t=\tilde{\sigma }_t\) for \(y=0\). Hence, from (75) and (77), we find that

$$\begin{aligned}&(\!( M_1z,M_2z)\!)+(\!(\tilde{M}_1\tilde{z},\tilde{M}_2\tilde{z})\!)\nonumber \\&\quad \ge C\int \!\!\!\!\int _{Q'}\left[ s^3\lambda ^4(\xi ^3|z|^2+\tilde{\xi }^3|\tilde{z}|^2)+s\lambda ^2(\xi |z_y|^2+\tilde{\xi }|\tilde{z}_y|^2)\right] \,\mathrm{d}y\,\mathrm{d}t\nonumber \\&\qquad -\,C\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\left[ s^3\lambda ^4(\xi ^3|z|^2+\tilde{\xi }^3|\tilde{z}|^2)\right. \nonumber \\&\qquad \left. +s\lambda ^2(\xi |z_y|^2+\tilde{\xi }|\tilde{z}_y|^2)\right] \,\mathrm{d}y\,\mathrm{d}t\nonumber \\&\qquad +\,\!\!\left. \int _0^T\!\!\left[ 2s\lambda ^2|\eta '|^2(\tilde{\xi }\tilde{z}\tilde{z}_y+\xi zz_y)-s\lambda \eta '(\tilde{\xi }\tilde{z}_y^2-\xi z_y^2)\right. \right. \nonumber \\&\qquad \left. \left. +\,\tilde{z}\tilde{z}_{ty}+zz_{ty}\right] \,\mathrm{d}t\right| _{y=0}\!\!\!. \end{aligned}$$
(78)

On the other hand, in view of the boundary conditions satisfied by z and \(\tilde{z}\), we see that

$$\begin{aligned} \left. \int _0^T[zz_{yt}+\tilde{z}\tilde{z}_{yt}]\mathrm{d}t\right| _{y=0}= & {} \left. \int _0^T\left[ s\lambda \xi _t(|\tilde{z}|^2-|z|^2)+\frac{1}{4}(|\tilde{z}|^2+|z|^2)_t\right. \right. \\&\left. \left. +\frac{1}{2}s\lambda \xi (|\tilde{z}|^2-|z|^2)_t\right] \mathrm{d}t\right| _{y=0}= 0,\\ \left. 2s\lambda ^2\!\!\int _0^T|\eta '|^2\xi [zz_y+\tilde{z}\tilde{z}_y]\,\mathrm{d}t\right| _{y=0}= & {} \left. 2s\lambda ^2\int _0^T|\eta '|^2\xi |z|^2\,\mathrm{d}t\right| _{y=0}\ge 0 \end{aligned}$$

and

$$\begin{aligned} -\left. s\lambda \int _0^T\eta '\xi [\tilde{z}_y^2-z_y^2]\,\mathrm{d}t\right| _{y=0}= & {} \left. 2s^2\lambda ^2\int _0^T\xi ^2 |z|^2\,\mathrm{d}t\right| _{y=0}\ge 0. \end{aligned}$$

As a consequence, from (74), (76) and (78), we deduce that

$$\begin{aligned}&\displaystyle \sum _{i=1}^{2}(\Vert M_{i}z\Vert ^{2}+\Vert \tilde{M}_{i}\tilde{z}\Vert ^{2})+\int \!\!\!\!\int _{Q^{\prime }}\left[ s\lambda ^{2}(\xi |z_{y} |^{2}+\tilde{\xi }|\tilde{z}_{y}|^{2})+s^{2}\lambda ^{4}(\xi ^{3}|z|^{2} +\tilde{\xi }^{3}|\tilde{z}|^{2})\right] \,\mathrm{d}y\,\mathrm{d}t\nonumber \\&\quad \le C\left( \Vert g_{1}\Vert ^{2}+\Vert \tilde{g}_{1}\Vert ^{2}+\int \!\!\!\!\int _{\omega _{0T}}\!\!\!\left[ s\lambda ^{2}(\xi |z_{y}|^{2}+\tilde{\xi }|\tilde{z}_{y}|^{2})+s^{3}\lambda ^{4}(\xi ^{3}|z|^{2}+\tilde{\xi } ^{3}|\tilde{z}|^{2})\right] \,\mathrm{d}y\,\mathrm{d}t\right) .\nonumber \\ \end{aligned}$$
(79)

Using (79) and the definitions of \(g_1\), \(\tilde{g}_1\), \(M_iz\) and \(\tilde{M}_i\tilde{z}\), we see that, for s and \(\lambda \) large enough,

$$\begin{aligned}&\int \!\!\!\!\int _{Q^{\prime }}\left[ s^{-1}\left[ \xi ^{-1}(|z_{t} |^{2}+|z_{yy}|^{2})+\tilde{\xi }^{-1}(|\tilde{z}_{t}|^{2}+|\tilde{z}_{yy} |^{2})\right] \right. \\&\qquad \left. +s\lambda ^{2}(\xi |z_{y}|^{2}+\tilde{\xi }|\tilde{z}_{y} |^{2})+\,s^{2}\lambda ^{4}(\xi ^{3}|z|^{2}+\tilde{\xi }^{3}|\tilde{z}|^{2})\right] \,\mathrm{d}y\,\mathrm{d}t\\&\quad \le C\left( \Vert \varrho ^{1/2}g_{0}\Vert ^{2}+\int \!\!\!\!\int _{\omega _{0T} }\!\!\!\left[ s\lambda ^{2}(\xi |z_{y}|^{2}+\tilde{\xi }|\tilde{z}_{y} |^{2})+s^{3}\lambda ^{4}(\xi ^{3}|z|^{2}+\tilde{\xi }^{3}|\tilde{z}|^{2})\right] \,\mathrm{d}y\,\mathrm{d}t\right) . \end{aligned}$$

From classical arguments, we can eliminate the terms with derivatives in the right-hand side of the last inequality and find that

$$\begin{aligned}&\int \!\!\!\!\int _{Q^{\prime }}\left[ s^{-1}\left[ \xi ^{-1}(|z_{t} |^{2}+|z_{yy}|^{2})+\tilde{\xi }^{-1}(|\tilde{z}_{t}|^{2}+|\tilde{z}_{yy} |^{2})\right] +s\lambda ^{2}(\xi |z_{y}|^{2}+\tilde{\xi }|\tilde{z}_{y} |^{2})\right. \\&\qquad \left. +\,s^{2}\lambda ^{4}(\xi ^{3}|z|^{2}+\tilde{\xi }^{3}|\tilde{z}|^{2})\right] \,\mathrm{d}y\,\mathrm{d}t\\&\quad \le C\left( \Vert \varrho ^{1/2}g_{0}\Vert ^{2}+s^{3}\lambda ^{4}\int \!\!\!\!\int _{\omega _{0T}}\left( \xi ^{3}|z|^{2}+\tilde{\xi }^{3}|\tilde{z}|^{2}\right) \,\mathrm{d}y\,\mathrm{d}t\right) . \end{aligned}$$

We then conclude the proof coming back to the original variable w and using the definition of \(g_0\) and the fact that \(\xi \le C\tilde{\xi }\le C\xi \). \(\Box \)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Araruna, F.D., Araújo, B.S.V. & Fernández-Cara, E. Stackelberg–Nash null controllability for some linear and semilinear degenerate parabolic equations. Math. Control Signals Syst. 30, 14 (2018). https://doi.org/10.1007/s00498-018-0220-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s00498-018-0220-6

Keywords

Mathematics Subject Classification

Navigation