Let us fix the initial condition
\((x_0, x_1) \in \mathbb {R}^2 \times L^2([-d,0];\mathbb {R})\). Let
\(T>0\), we introduce the space
$$\begin{aligned} S_T:=\{y \in C([0,T];\mathbb {R}):y(0)=x_0\}, \end{aligned}$$
endowed with the sup norm
$$\begin{aligned} \Vert y\Vert _{S_T} = \sup _{t \in [0,T]}|y(t)|. \end{aligned}$$
We consider the space
\(L^p(\Omega ;S_T)\),
\(p \ge 2\), endowed with the norm
$$\begin{aligned} \Vert y\Vert _{L^p(\Omega ;S_T)} = \left( \mathbb {E}\left[ \Vert y\Vert ^p_{S_T}\right] \right) ^{\frac{1}{p}}= \left( \mathbb {E}\left[ \sup _{t \in [0,T]}|y(t)|^p\right] \right) ^{\frac{1}{p}}. \end{aligned}$$
In the sequel we will denote by
\(p':=\frac{p}{p-1}\) the conjugate exponent to
p and by
\(p^*:=\frac{p}{p-2}\) the conjugate exponent to
\(\frac{p}{2}\). Given
\(y \in L^p(\Omega ;S_T)\), let
$$\begin{aligned} F(y)(t)&:= x_0+ \mu _y \int _0^t y(r)\,\textrm{d}r + \int _0^t \mathcal {L}({\bar{y}}^{x_1})\, \textrm{d}r+ \int _0^t y(r)\sigma _y \cdot \textrm{d}Z(r)\nonumber \\&\quad \quad + \int _0^t\mathcal {G}({\bar{y}}^{x_1})\cdot \textrm{d}Z(r), \qquad 0 \le t \le T. \end{aligned}$$
(48)
Here
\(\mathcal {L}\) and
\(\mathcal {G}\) are the continuous linear operators introduced in Lemma
5.1 and
\({\bar{y}}^{x_1} \in L^p(\Omega ;L^2([-d,T];\mathbb {R}) )\) is defined as follows:
$$\begin{aligned} \bar{y}^{x_1}(t) = {\left\{ \begin{array}{ll} x_1(t), &{} \text {if } -d\le t< 0; \\ y(t), &{} \text {if } \ \ 0\le t\le T. \end{array}\right. } \end{aligned}$$
(49)
We aim at proving that
F maps
\(L^p(\Omega ;S_T)\) into itself for any
\(p\ge 2\) and that it is a contraction on the same space when
\(p>4\).
Let us start by proving that
F maps
\(L^p(\Omega , S_T)\),
\(p\ge 2\), into itself. We write
$$\begin{aligned}&\Vert F(y)\Vert _{L^p(\Omega ;S_T)} \le |x_0|+ |\mu _y|\left\| \int _0^{\cdot }y(r)\, \textrm{d}r\right\| _{L^p(\Omega ;S_T)} \nonumber \\&\quad + \left\| \int _0^{\cdot } \mathcal {L}({\bar{y}}^{x_1})\, \textrm{d}r \right\| _{L^p(\Omega ;S_T)} +\left\| \int _0^{\cdot } y(r)\sigma _y \cdot \, \textrm{d}Z(r)\right\| _{L^p(\Omega ;S_T)}\nonumber \\&\quad + \left\| \int _0^{\cdot } \mathcal {G}({\bar{y}}^{x_1}) \cdot \, \textrm{d}Z(r)\right\| _{L^p(\Omega ;S_T)}. \end{aligned}$$
(50)
The boundedness of the terms that appears in the r.h.s. of (
50), except the last one, can be proved following the lines of [
6, Proposition B.2]. We estimate the last term in the r.h.s. of (
50) by means of the Burkholder-Davies-Gundy inequality
$$\begin{aligned} \left\| \int _0^{\cdot } \mathcal {G}({\bar{y}}^{x_1}) \cdot \, \textrm{d}Z(r)\right\| _{L^p(\Omega ;S_T)}^p&= \mathbb {E} \left[ \sup _{t \in [0,T]}\left| \int _0^t \mathcal {G}({\bar{y}}^{x_1})\cdot \textrm{d}Z(r)\right| ^p \right] \\&\lesssim \mathbb {E} \left[ \left| \int _0^T\Vert \mathcal {G}({\bar{y}}^{x_1})\Vert ^2\, \textrm{d}r\right| ^{\frac{p}{2}} \right] \\&= \mathbb {E} \left[ \Vert \mathcal {G}({\bar{y}}^{x_1})\Vert ^p_{L^2([0,T];\mathbb {R}^n)}\right] \lesssim \mathbb {E} \left[ \Vert {\bar{y}}^{x_1}\Vert ^p_{L^2([-d,T];\mathbb {R})}\right] \\&=\Vert {\bar{y}}^{x_1}\Vert ^p_{L^p(\Omega ;L^2([-d,T];\mathbb {R}))} < \infty , \end{aligned}$$
where in the last inequality we exploited (
47) of Lemma
5.1.
Given
\(y, z \in L^p(\Omega ;S_T)\), from (
48) and (
51), we have
$$\begin{aligned} \Vert F(z)-F(y)\Vert _{\alpha }^p \lesssim _p&\mathbb {E} \left[ \sup _{t \in [0,T]} e^{-p\alpha t}\left( |\mu _y|\left| \int _0^{t}(z(r)-y(r))\textrm{d}r\right| ^p+ \left| \int _0^{t} \mathcal {L}({\bar{z}}^{x_1}-{\bar{y}}^{x_1})\, \textrm{d}r \right| ^p \right) \right] \nonumber \\&+\mathbb {E}\left[ \sup _{t \in [0,T]} e^{-p\alpha t}\left| \int _0^{t} (z(r)-y(r))\sigma _y \cdot \, \textrm{d}Z(r)\right| ^p\right] \nonumber \\&+\mathbb {E}\left[ \sup _{t \in [0,T]} e^{-p\alpha t}\left| \int _0^{t} \mathcal {G}({\bar{z}}^{x_1}-{\bar{y}}^{x_1})\cdot \, \textrm{d}Z(r)\right| ^p\right] . \end{aligned}$$
(52)
We can estimate the first three terms in the r.h.s. of (
52) proceeding as in [
6, Proposition B.2]
4. For the first term we obtain
$$\begin{aligned}&\mathbb {E}\left[ \sup _{t \in [0,T]} e^{-p\alpha t} |\mu _y|\left| \int _0^{t}(z(r)-y(r))\textrm{d}r\right| ^p\right] \nonumber \\&\quad \le |\mu _y|T\left( \frac{1}{\alpha p'}\right) ^{\frac{p}{p'}}\Vert z-y\Vert ^p_{\alpha } \lesssim _{\mu _y,T, p} C_1(\alpha )\Vert z-y\Vert ^p_{\alpha }. \end{aligned}$$
(53)
For the second term we get
$$\begin{aligned}{} & {} \mathbb {E} \left[ \sup _{t \in [0,T]} e^{-p\alpha t} \left| \int _0^{t} \mathcal {L}({\bar{z}}^{x_1}-{\bar{y}}^{x_1})\, \textrm{d}r \right| ^p \right] \nonumber \\{} & {} \quad \quad \le \left( \frac{|\phi |([-d,0])}{\alpha p'} \right) ^{\frac{p}{p'}}T |\phi |([-d,0])\Vert z-y\Vert ^p_{\alpha } \lesssim _{|\phi |, p, T}C_2(\alpha )\Vert z-y\Vert ^p_{\alpha }. \end{aligned}$$
(54)
For the third term, by means of the so called factorization method (see e.g. [?, Section 5.3]), for a given
\(\delta \in \left( \frac{1}{p}, \frac{1}{2}\right) \)5, we have
$$\begin{aligned} \mathbb {E}&\left[ \sup _{t \in [0,T]} e^{-p\alpha t}\left| \int _0^{t} (z(r)-y(r))\sigma _y \cdot \, \textrm{d}Z(r)\right| ^p\right] \nonumber \\&\lesssim _{p, \delta } \left( \int _0^T u^{p'(\delta -1)}e^{-p'\alpha u}\, \textrm{d}u \right) ^{\frac{p}{p'}}T \Vert \sigma _y\Vert ^p\left( \sup _{u \in [0,T] }\int _0^u(u-r)^{-2\delta }e^{-2\alpha (u-r)}\, \textrm{d}r\right) ^{\frac{p}{2}}\nonumber \\&\quad \Vert z-y\Vert _{\alpha }^p \lesssim _{p,\delta , T, \Vert \sigma _y\Vert } C_3(\alpha )\Vert z-y\Vert ^p_{\alpha }. \end{aligned}$$
(55)
Let us now come to the estimate of the fourth term in (
52). Exploiting the factorization method, for
\(\eta \in \left( \frac{1}{p}, \frac{p-2}{2p}\right) \)6 we rewrite that stochastic integral as follows
$$\begin{aligned} \int _0^t \mathcal {G}({\bar{z}}^{x_1}-{\bar{y}}^{x_1}) \cdot \,{ \mathrm d}Z(r)=c_{\eta }\int _0^t (t-u)^{\eta -1} Y(u)\, \textrm{d}u, \end{aligned}$$
with
$$\begin{aligned} \frac{1}{c_{\eta }}:= \int _r^t (t-u)^{\eta -1}(u-r)^{-\eta }\, \textrm{d}u = \frac{\pi }{\sin (\pi \eta )}, \end{aligned}$$
and
$$\begin{aligned} Y(u)=\int _0^u (u-r)^{-\eta }\mathcal {G}({\bar{z}}^{x_1}-{\bar{y}}^{x_1}) \cdot \textrm{d}Z(r). \end{aligned}$$
Thanks to the Hölder inequality we estimate
$$\begin{aligned} e^{-\alpha t}\left| \int _0^t \mathcal {G}({\bar{z}}^{x_1}-{\bar{y}}^{x_1})\cdot \textrm{d}Z(r)\right|&= c_{\eta } e^{-\alpha t}\left| \int _0^t (t-u)^{\eta -1}Y(u)\, \textrm{d}u \right| \\&= c_{\eta } \left| \int _0^t e^{-\alpha (t-u)}(t-u)^{\eta -1}e^{-\alpha u}Y(u)\, \textrm{d}u \right| \\&\le c_{\eta }\left( \int _0^te^{-\alpha p'(t-u)}(t-u)^{p'(\eta -1)}\, \textrm{d}u\right) ^{\frac{1}{p'}} \\&\quad \left( \int _0^te^{-\alpha p u}|Y(u)|^p\, \textrm{d}u\right) ^{\frac{1}{p}}. \end{aligned}$$
Therefore we obtain
$$\begin{aligned} \mathbb {E}{} & {} \left[ \sup _{t \in [0,T]}e^{-\alpha p t}\left| \int _0^t \mathcal {G}({\bar{z}}^{x_1}-{\bar{y}}^{x_1}) \cdot \textrm{d}Z(r) \right| ^p \right] \\{} & {} \le c_{\eta }^p \mathbb {E}\left[ \sup _{t \in [0,T]} \left( \int _0^te^{-\alpha p'(t-u)}(t-u)^{p'(\eta -1)}\, \textrm{d}u\right) ^{\frac{p}{p'}} \left( \int _0^te^{-\alpha p u}|Y(u)|^p\, \textrm{d}u\right) \right] \\{} & {} \le c^p_{\eta } \left( \int _0^Te^{-\alpha p'u}u^{p'(\eta -1)}\, \textrm{d}u\right) ^{\frac{p}{p'}} \int _0^T e^{-\alpha pu} \mathbb {E} \left[ |Y(u)|^p\right] \, \textrm{d}u. \end{aligned}$$
Now, recalling the definition of
\(\mathcal {G}\) and that, when
\(r<d\),
\({\bar{z}}^{x_1}_r(s)-{\bar{y}}^{x_1}_r(s)=0\) for
\(s \in [-d,-r)\) (see (
49)), by means of the Burkholder-Davis-Gundy (BDG) and the Hölder (H) inequalities, we obtain for all
\(u \in [0,T]\),
$$\begin{aligned} e^{-\alpha pu}{} & {} \mathbb {E}\left[ |Y(u)|^p\right] = e^{-\alpha pu} \mathbb {E}\left[ \left| \int _0^u (u-r)^{-\eta }\mathcal {G}({\bar{z}}^{x_1}-{\bar{y}}^{x_1}) \cdot \textrm{d}Z(r)\right| ^p \right] \\{} & {} \quad \overset{BDG}{\lesssim _p}\ e^{-\alpha up} \mathbb {E}\left[ \left| \int _0^u (u-r)^{-2\eta }\Vert \mathcal {G}({\bar{z}}^{x_1}-{\bar{y}}^{x_1})\Vert ^2\, \textrm{d}r\right| ^{\frac{p}{2}}\right] \\{} & {} \quad \quad =e^{-\alpha up}\mathbb {E}\left[ \left| \int _0^u (u-r)^{-2\eta } \sum _{i=1}^n \left| \int _{-d}^0({\bar{z}}_r^{x_1}-{\bar{y}}_r^{x_1})(s)\, \varphi _i(\textrm{d}s) \right| ^2 \, \textrm{d}r\right| ^{\frac{p}{2}}\right] \\{} & {} \quad \quad =e^{-\alpha up}\mathbb {E}\left[ \left| \int _0^u (u-r)^{-2\eta } \sum _{i=1}^n \left| \int _{-d\vee -r}^0\left( (z(r+s)-y(r+s)\right) \, \varphi _i(\textrm{d}s) \right| ^2 \, \textrm{d}r\right| ^{\frac{p}{2}}\right] \\{} & {} \quad \quad =\mathbb {E}\left[ \left| \int _0^u (u-r)^{-2\eta }e^{-2\alpha (u-r-s)}e^{-2\alpha (r+s)}\sum _{i=1}^n \right. \right. \\{} & {} \qquad \left. \left. \left| \int _{-d\vee -r}^0\left( z(r+s)-y(r+s)\right) \, \varphi _i(\textrm{d}s) \right| ^2 \, \textrm{d}r\right| ^{\frac{p}{2}}\right] \\{} & {} \quad \quad \overset{H}{\le }\mathbb {E}\left[ \left| \int _0^u (u-r)^{-2\eta }e^{-2\alpha (u-r-s)}e^{-2\alpha (r+s)}\sum _{i=1}^n |\varphi _i|([-d,0]) \right. \right. \\{} & {} \quad \left. \left. \int _{-d\vee -r}^0\left| (z(r+s)-y(r+s)\right| ^2\, \varphi _i(\textrm{d}s) \, \textrm{d}r\right| ^{\frac{p}{2}}\right] \\{} & {} \quad \overset{H}{\le }\left( \sum _{i=1}^n \left( |\varphi _i|([-d,0]) \right) ^{p^*}\int _0^u\int _{-d \vee -r}^0(u-r)^{-2p^*\eta }e^{-2\alpha p^*(u-r-s)}\, \varphi _i(\textrm{d}s)\, \textrm{d}r\right) ^{\frac{p}{2p^*}} \\{} & {} \qquad \quad \ \mathbb {E} \left[ \sum _{i=1}^n\int _0^u \int _{-d \vee -r}^0 e^{-\alpha p(r+s)}|z(r+s)-y(r+s)|^p\, \varphi _i(\textrm{d}s)\, \textrm{d}r \right] \\{} & {} \quad \quad \le \left( \sum _{i=1}^n \left( |\varphi _i|([-d,0]) \right) ^{p^*}\int _0^u\int _{-d \vee -r}^0(u-r)^{-2p^*\eta }e^{-2\alpha p^*(u-r-s)}\, \varphi _i(\textrm{d}s)\, \textrm{d}r\right) ^{\frac{p}{2p^*}} \\{} & {} \qquad \sum _{i=1}^n\int _0^u \int _{-d \vee -r}^0\mathbb {E} \left[ \sup _{(r+s) \in [0,u]}\left( e^{-\alpha p(r+s)}|z(r+s)-y(r+s)|^p\right) \right] \, \varphi _i(\textrm{d}s)\, \textrm{d}r \\{} & {} \quad \quad \le \left( \sum _{i=1}^n \left( |\varphi _i|([-d,0]) \right) ^{p^*}\int _0^u\int _{-d \vee -r}^0(u-r)^{-2p^*\eta }e^{-2\alpha p^*(u-r-s)}\, \varphi _i(\textrm{d}s)\, \textrm{d}r\right) ^{\frac{p}{2p^*}} \\{} & {} \qquad \quad \ u \sum _{i=1}^n |\varphi _i|([-d,0]) \mathbb {E} \left[ \sup _{(r+s) \in [0,u]}\left( e^{-\alpha p(r+s)}|z(r+s)-y(r+s)|^p\right) \right] . \end{aligned}$$
Therefore,
$$\begin{aligned} \int _0^T{} & {} e^{-\alpha pu}\mathbb {E}\left[ |Y(u)|^p\right] \, \textrm{d}u \\{} & {} \lesssim _{|\varphi _i|,p} \int _0^Tu \left( \sum _{i=1}^n \int _0^u\int _{-d \vee -r}^0(u-r)^{-2p^*\eta }e^{-2\alpha p^*(u-r-s)}\, \varphi _i(\textrm{d}s)\, \textrm{d}r\right) ^{\frac{p}{2p^*}} \\{} & {} \qquad \quad \mathbb {E} \left[ \sup _{(r+s) \in [0,u]}\left( e^{-\alpha p(r+s)}|z(r+s)-y(r+s)|^p\right) \right] \, \textrm{d}u \\{} & {} \lesssim _{|\varphi _i|,p} \int _0^Tu \left( \sum _{i=1}^n \int _0^u\int _{-d \vee -r}^0(u-r)^{-2p^*\eta }e^{-2\alpha p^*(u-r-s)}\, \varphi _i(\textrm{d}s)\, \textrm{d}r\right) ^{\frac{p}{2p^*}} \\{} & {} \qquad \quad \ \mathbb {E} \left[ \sup _{(r+s) \in [0,T]}\left( e^{-\alpha p(r+s)}|z(r+s)-y(r+s)|^p\right) \right] \, \textrm{d}u \\{} & {} =\left( \int _0^Tu \left( \sum _{i=1}^n \int _0^u\int _{-d \vee -r}^0(u-r)^{-2p^*\eta }e^{-2\alpha p^*(u-r-s)}\, \varphi _i(\textrm{d}s)\, \textrm{d}r\right) ^{\frac{p}{2p^*}} \, \textrm{d}u \right) \ \Vert z-y\Vert _{\alpha }^p \\{} & {} \lesssim _{|\varphi _i|,T,p} \left( \int _0^Tr^{-2p^*\eta }e^{-2\alpha p^*r} \, \textrm{d}r\right) ^{\frac{p}{2p^*}} \Vert z-y\Vert _{\alpha }^p \end{aligned}$$
where the last inequality is obtained as follows:
$$\begin{aligned} \int _0^T&u \left( \sum _{i=1}^n \int _0^u\int _{-d \vee -r}^0(u-r)^{-2p^*\eta }e^{-2\alpha p^*(u-r-s)}\, \varphi _i(\textrm{d}s)\, \textrm{d}r\right) ^{\frac{p}{2p^*}} \, \textrm{d}u \\&\le T \int _0^T \left( \sum _{i=1}^n \int _0^u\int _{-d \vee -r}^0(u-r)^{-2p^*\eta }e^{-2\alpha p^*(u-r)} e^{2\alpha p^*s}\, \varphi _i(\textrm{d}s)\, \textrm{d}r\right) ^{\frac{p}{2p^*}} \, \textrm{d}u \\&\le T \int _0^T \left( \sum _{i=1}^n \int _0^u\int _{-d \vee -r}^0(u-r)^{-2p^*\eta }e^{-2\alpha p^*(u-r)} \, \varphi _i(\textrm{d}s)\, \textrm{d}r\right) ^{\frac{p}{2p^*}} \, \textrm{d}u \\&\le T \sup _{u \in [0,T]} \left( \sum _{i=1}^n |\varphi _i|([-d,0])\int _0^u(u-r)^{-2p^*\eta }e^{-2\alpha p^*(u-r)} \, \textrm{d}r\right) ^{\frac{p}{2p^*}} \\&\lesssim _{|\varphi _i|, T}\left( \int _0^Tr^{-2p^*\eta }e^{-2\alpha p^*r} \, \textrm{d}r\right) ^{\frac{p}{2p^*}}. \end{aligned}$$
Putting together the above estimates we obtain
$$\begin{aligned} \mathbb {E}&\left[ \sup _{t \in [0,T]}e^{-\alpha p t}\left| \int _0^t \mathcal {G}({\bar{z}}^{x_1}-{\bar{y}}^{x_1}) \cdot \textrm{d}Z(r) \right| ^p \right] \nonumber \\&\lesssim _{T, |\varphi _i|,\eta ,p} \left( \int _0^Te^{-\alpha p'u}u^{p'(\eta -1)}\, \textrm{d}u\right) ^{\frac{p}{p'}} \left( \int _0^Tr^{-2p^*\eta }e^{-2\alpha p^*r} \, \textrm{d}r\right) ^{\frac{p}{2p^*}} \Vert z-y\Vert _{\alpha }^p \nonumber \\&\lesssim _{T, |\varphi _i|, \eta ,p} C_4(\alpha ) \Vert z-y\Vert _{\alpha }^p, \end{aligned}$$
(56)
Finally, from (
53), (
54), (
55) and (
56) we infer
$$\begin{aligned} \Vert F(z)-F(y)\Vert ^p_{\alpha } \lesssim _{\mu _y,T,p,|\phi |,|\varphi _i|, \Vert \sigma _y\Vert , \delta , \eta } \sum _{i=1}^4 C_i(\alpha ) \Vert z-y\Vert _{\alpha }^p, \end{aligned}$$
where
\(C_i(\alpha ) \rightarrow 0\) as
\(\alpha \rightarrow \infty \), for
\(i=1,\ldots ,4\). So, by taking
\(\alpha >0\) sufficiently large, this proves that
F is a contraction, thus there exists a unique fixed point of it. In this way we prove the existence and uniqueness of the solution in the space
\(L^p(\Omega ,S_T)\) for
\(p >4\). Since, for such
p,
\(L^p(\Omega ,S_T)\subset L^2(\Omega ,S_T)\), such solution also belongs to
\(L^2(\Omega ,S_T)\). To get uniqueness in the space
\(L^2(\Omega ,C([0,T];\mathbb {R}))\) one can take two solutions
y and
\({{\tilde{y}}}\) in this space and take their difference. Using the fact that both are fixed points of
F, by means of the Gronwall Lemma one gets
\(\sup _{t\in [0,T]} \mathbb {E}\left[ |y(t)-{{\tilde{y}}}(t)|^2\right] =0\) and this concludes the proof.
\(\square \)