1: \begin{proof}[Proof of Proposition~\ref{prn:HD07_2}]

2: It suffices to check the three steps mentioned in the strategy. 

3: 

4: We first check that $\dot{\Psi}_0(s)=2\Psi(s)$ for each $s \in (0,1)$.  Formal calculation shows that

5: \begin{align*}

6:    \dot{\Psi}_0(s)&=\int_X \left( \frac{d}{ds} u^2(1-s,x) \right) p(s,x,y) d\mu_x  + \int_X u^2(1-s,x) \Delta p(s,x,y) d\mu_x\\

7:       &=\int_X \left\{  -2u(x,1-s) \dot{u}(x,1-s) \right\} p(s,x,y) d\mu_x  + \int_X u^2(1-s,x)  \Delta p(s,x,y) d\mu_x.       

8: \end{align*}

9: By the boundedness of $u, \dot{u}$ and exponential decay of $p$ and $\Delta p$, the above formal calculation is in fact rigorous for each $s \in (0,1)$.

10: Moreover, $u^2(1-s,\cdot) \in L^{\infty}(X) \cap N^{1,2}(X)$.  Since $\Delta p$ has exponential decay, similar to Lemma~\ref{lma:HH09_2}, one can have 

11: integration by parts to obtain

12: \begin{align*}

13:   \int_X u^2(1-s,x) \Delta p(s,x,y) d\mu_x=\int_X p(s,x,y) \Delta u^2(1-s,x) d\mu_x. 

14: \end{align*}

15: Therefore, we have

16: \begin{align*}   

17:        \dot{\Psi}_0(s)

18:       &=\int_X \left\{-2u(x,1-s) \dot{u}(x,1-s) \right\} p(s,x,y) d\mu_x \\

19:       &\qquad + \int_X \left\{ 2u(x,1-s) \Delta u(x,1-s) +2|\nabla u|^2(x,1-s) \right\} p(s,x,y) d\mu_x\\

20:       &=2 \int_X |\nabla u|^2 (x,1-s) p(s,x,y) d\mu_x

21:          =2\Psi(s).

22: \end{align*}

23: So we checked the first step.  

24: However, the second step follows from Lemma~\ref{lma:HH08_1},  the third step follows from Lemma~\ref{lma:HH12_1}. 

25: Therefore, the proof of Proposition~\ref{prn:HD07_2} is complete.  

26: \end{proof}

27: