- Review
- Open access
- Published:
Quadratic covariations for the solution to a stochastic heat equation with space-time white noise
Advances in Difference Equations volume 2020, Article number: 254 (2020)
Abstract
Let \(u(t,x)\) be the solution to a stochastic heat equation
with initial condition \(u(0,x)\equiv 0\), where Ẋ is a space-time white noise. This paper is an attempt to study stochastic analysis questions of the solution \(u(t,x)\). In fact, it is well known that the solution is a Gaussian process such that the process \(t\mapsto u(t,x)\) is a bi-fractional Brownian motion with Hurst indices \(H=K=\frac{1}{2}\) for every real number x. However, the many properties of the process \(x\mapsto u(\cdot ,x)\) are unknown. In this paper we consider the generalized quadratic covariations of the two processes \(x\mapsto u(\cdot ,x),t\mapsto u(t,\cdot )\). We show that \(x\mapsto u(\cdot ,x)\) admits a nontrivial finite quadratic variation and the forward integral of some adapted processes with respect to it coincides with “Itô’s integral”, but it is not a semimartingale. Moreover, some generalized Itô formulas and Bouleau–Yor identities are introduced.
1 Introduction
Let \(u(t,x)\) denote the solution to the stochastic heat equation
with initial condition \(u(0,x)\equiv 0\), where Ẋ is a space-time white noise on \([0,\infty )\times {\mathbb{R}}^{d}\), which are generalized Gaussian processes with covariance given by
Clearly, we have
where \(p(t,x)=\frac{1}{\sqrt{2\pi t}}e^{-\frac{|x|^{2}}{2t}}\) is the heat kernel. Then the processes \((t,x)\mapsto u(t,x)\), \(t\mapsto u(t,\cdot )\) and \(x\mapsto u(\cdot ,x)\) are Gaussian. Mueller and Tribe [19] were first to find that
for all \(x,y\in {\mathbb{R}}^{d}\), as t tends to infinity, if the initial value \(u(0,x)=0\). By taking a special initial value (a two-sided \({\mathbb{R}}^{d}\)-valued Brownian motion), the authors studied the stationary pinned string and hitting probabilities of a random string. When \(d=1\) Swanson [30] has showed that (see also Pospisil and Tribe [25])
and the process \(t\mapsto u(t,x)\) has a nontrivial quartic variation. This shows that, for every \(x\in {\mathbb{R}}\), the process \(t\mapsto u(t,x)\) coincides with the bi-fractional Brownian motion and it is not a semimartingale, so a stochastic integral with respect to the process \(t\mapsto u(t,x)\) cannot be defined in the classical Itô sense. Some surveys and complete literature for bi-fractional Brownian motion can be found in Houdré and Villa [13], Lei and Nualart [16], Russo and Tudor [27], Tudor and Xiao [33] and Yan et al. [35], and the references therein. Moreover, for more general parabolic SPDEs, many authors studied the regularity results and stochastic calculus with respect to their solutions. We mention the work of Balan and Kim [2], Da Prato et al. [4], Deya and Tindel [5], Gradinaru et al. [12], Lanconelli [14, 15], León and Tindel et al. [17], Nualart and Vuillermot [21], Ouahhabi and Tudor [22], Pardoux [23], Pospisil and Tribe [25], Torres et al. [31], Tudor [32], Tudor and Xiao [34], Zambotti [38], and the references therein.
In this paper, as an attempt we study stochastic analysis questions of the solution process \(u=\{u(t,x),t\in [0,T],x\in {\mathbb{R}}\}\) of (1.1) with \(u(0,x)=0\) associated with quadratic variation. We shall see (in Sect. 3) that the process \(x\mapsto u(\cdot ,x)\) admits a nontrivial finite quadratic variation in any finite interval, and, moreover, we shall also see (in Sect. 4) that the forward integral of some adapted processes with respect to \(x\mapsto u(\cdot ,x)\) coincides with “Itô’s integral”. On the other hand, as a noise, the stochastic process \(u=\{u(t,x), t\geq 0,x\in {\mathbb{R}}\}\) is very rough in time and it is not white in space. However, the process \(x\mapsto u(\cdot ,x)\) admits some characteristics similar to Brownian motion. These results, together with the work of Mueller and Tribe [19], Pospisil and Tribe [25], and Swanson [30] point out that the process \(u=\{u(t,x)\}\) as a noise admits the next special structures:
It is very rough in time and similar to fractional Brownian motion with Hurst index \(H=\frac{1}{4}\), but it has not stationary increments.
It is not white in space, but its quadratic variation coincides with the classical Brownian motion and it is not self-similar.
The process in space variable is not a semimartingale, but the forward integral of some adapted processes with respect to the process in space variable coincides with “Itô’s integral”.
The process \(u=\{u(t,x)\}\) admits a simple representation via a Wiener integral with respect to the Brownian sheet.
Though the process \(u=\{u(t,x)\}\) is Gaussian, as a noise, its time and space parts are farraginous. We cannot decompose its covariance as the product of two independent parts. This is very different from fractional noise and white noise. In fact, we have
$$\begin{aligned} Eu(t,x)u(s,y)=\frac{1}{\sqrt{2\pi }} \int _{0}^{s}\frac{1}{\sqrt{t+s-2r}} \exp \biggl\{ - \frac{(x-y)^{2}}{2(t+s-2r)} \biggr\} \,dr \end{aligned}$$for all \(t\geq s>0\) and \(x,y\in {\mathbb{R}}\).
Therefore, it seems interesting to study the integrals
and some related stochastic (partial) differential equations. For example, one can consider the following “iterated” stochastic partial differential equations:
where \(u^{0}\) is a space-time white noise. Of course, one can also consider some sample path properties and singular integrals
with \(t\geq 0\) and \(I_{x}=[0,x]\) for \(x\geq 0\) and \(I_{x}=[x,0]\) for \(x\leq 0\). We will carry out these projects in some forthcoming work. In the present paper our objects are to study the quadratic covariations of \(x\mapsto u(\cdot ,x)\) and \(t\mapsto u(t,\cdot )\), and introduce some generalized Itô formulas associated with \(\{u(\cdot ,x),x\in {\mathbb{R}}\}\) and \(\{u(t,\cdot ),t\geq 0\}\), respectively, and moreover we also consider their local times and Bouleau–Yor’s identities.
To expound our aim, let us start with a basic definition. Let \(u=\{u(t,x),t\in [0,T],x\in {\mathbb{R}}\}\) be the solution process of (1.1) with \(u(0,x)=0\) and denote
and
for all \(t,s>0\) and \(z\geq 0\). An elementary calculation can show that
for all \(t,s>0\) and \(x,y\in {\mathbb{R}}\). This simple estimate inspires us to consider the following limits:
and
for all \(t\geq 0\) and \(x\in {\mathbb{R}}\). That is,
for all \(t\geq 0\) and \(x\in {\mathbb{R}}\). Thus, the next definition is natural.
Definition 1.1
Denote \(B:=\{B_{t}:=u(t,\cdot ),t\geq 0\}\) and \(W:=\{W_{x}:=u(\cdot ,x),x\in {\mathbb{R}}\}\). Let \(I_{x}=[0,x]\) for \(x\geq 0\) and \(I_{x}=[x,0]\) for \(x\leq 0\). Define the integrals
for all \(t\geq 0,x\in {\mathbb{R}},\varepsilon ,\delta >0\), where f is a measurable function on \({\mathbb{R}}\). The limits \(\lim_{\delta \to 0}I_{\delta }^{1}(f, t,x)\) and \(\lim_{\varepsilon \to 0}I_{\varepsilon }^{2}(f,t,x)\) are called the partial quadratic covariations (PQC, in short) in space and in time, respectively, of \(f(u)\) and u, provided these limits exist in probability. We denote them by \([f(W),W]^{(SQ)}_{x}\) and \([f(B),B]^{(TQ)}_{t}\), respectively.
Clearly, we have (see Sect. 3)
and \([W,W]^{(SQ)}_{x}=|x|\) for all \(f\in C^{1}({\mathbb{R}}),t>0,x\in {\mathbb{R}}\). We also have (see Sect. 6)
and \([B,B]^{(TQ)}_{t}=\sqrt{\frac{2}{\pi }t}\) for all \(f\in C^{1}({\mathbb{R}}),t\geq 0,x\in {\mathbb{R}}\). These expressions say that the process \(W=\{W_{x}=u(\cdot ,x),x\in {\mathbb{R}}\}\) admits a nontrivial finite quadratic variation in any finite interval \(I_{x}\). This is also a main motivation to study the solution of (1.1). More work for stochastic calculus with respect to a continuous finite quadratic variation process can be found in Errami and Russo [7] and Russo and Vallois [29].
This paper is organized as follows. In Sect. 2, we establish some technical estimates associated with the solution of (1.1), and as some applications we introduce Wiener integrals with respect to the two processes \(B=\{B_{t}=u(t,\cdot ),t\geq 0\}\) and \(W=\{W_{x}=u(\cdot ,x),x\in {\mathbb{R}}\}\), respectively. In Sect. 3 we show that the quadratic variation \([W,W]^{(SQ)}\) exists in \(L^{2}(\varOmega )\) and equals \(|x|\) in every finite interval \(I_{x}\). For a given \(t>0\), by estimating in \(L^{2}\)
and
for all \(\varepsilon >0\), respectively, we construct a Banach space \({\mathscr{H}}_{t}\) of measurable functions such that the PQC \([f(W),W]^{(SQ)}\) in space exists in \(L^{2}(\varOmega )\) for all \(f\in {\mathscr{H}}_{t}\), and in particular we have
provided \(f\in C^{1}({\mathbb{R}})\). In Sect. 4, as an application of Sect. 3, we show that the Itô’s formula
holds for all \(t>0,x\in {\mathbb{R}}\), where the integral \(\int _{I_{x}}f'(W_{y})\delta W_{y}\) denotes the Skorohod integral, F is an absolutely continuous function with the derivative \(F'=f\in {\mathscr{H}}_{t}\). In order to obtain the above Itô formula, we first introduce a standard Itô type formula
for all \(F\in C^{2}({\mathbb{R}})\) satisfying some suitable conditions. It is important to note that the Gaussian process \(W=\{W_{x}=u(\cdot ,x),x\in {\mathbb{R}}\}\) does not satisfy the condition in Alós et al. [1] since
for all \(t\geq 0\) and \(x\in {\mathbb{R}}\). We need to give the proof of Eq. (1.6). Moreover, we also show that the forward integral (see Russo–Vallois [29])
coincides with the Skorohod integral \(\int _{I_{x}}f(W_{y})\delta W_{y}\), if f satisfies the growth condition
with \(0\leq \beta <\frac{\sqrt{\pi }}{4\sqrt{t}}\), where the notation ucplim denotes the uniform convergence in probability on each compact interval. This is very similar to Brownian motion, but the process \(W=\{W_{x}=u(\cdot ,x),x\in {\mathbb{R}}\}\) is not a semimartingale. In Sect. 5 we consider some questions associated with the local time
of the process \(W=\{W_{x}=u(\cdot ,x),x\geq 0\}\). In particular, we show that the Bouleau–Yor type identity
holds for all \(f\in {\mathscr{H}}_{t}\). In Sect. 6 we consider some analysis questions associated with the quadratic covariation of the process \(B=\{B_{t}=u(t,\cdot ),t\geq 0\}\).
2 Some basic estimates
In this section we will establish divergence integral and some technical estimates associated with the solution
where \(p(t,x)=\frac{1}{\sqrt{2\pi t}}e^{-\frac{x^{2}}{2t}}\) is the heat kernel. Recall that the temporal process \(B=\{B_{t}=u(t,\cdot ),t\geq 0\}\) and spatial process \(W=\{W_{x}=u(\cdot ,x),x\in {\mathbb{R}}\}\). As mentioned earlier, the temporal process \(B=\{B_{t}\}\) is neither a semimartingale nor a Markov process. Moreover, from the next discussion we shall also see that the spatial process \(W=\{W_{x}\}\) is neither a semimartingale nor a Markov process. However, as two Gaussian processes, one can develop the stochastic calculus of variations with respect to them. We refer to Alós et al. [1], Nualart [20] and the references therein for more details of stochastic calculus of Gaussian process.
As usual, we assume that \({\mathcal{H}}_{\vartriangle }\) and \({\mathcal{H}}_{\triangledown }\) are the reproducing kernel Hilbert spaces associated with \(W=\{W_{x}=u(\cdot ,x),x\in [-M,M]\}\) and \(B=\{B_{t}=u(t,\cdot ),t\in [0,T]\}\), respectively. Then the Wiener integrals
and
exist as two Gaussian random variables, and
for \(\varphi _{1},\varphi _{2}\in {\mathcal{H}}_{\vartriangle }\) and
for \(\psi _{1},\psi _{2}\in {\mathcal{H}}_{\triangledown }\). Clearly, for \(t,s>0\) and \(x,y\in {\mathbb{R}}\) we have
and
Denote by \({\mathcal{S}}_{\vartriangle }\) and \({\mathcal{S}}_{\triangledown }\) the sets of smooth functionals of the form
and
where \(f\in C^{\infty }_{b}({\mathbb{R}}^{n})\) (f and all their derivatives are bounded), \(\varphi _{i}\in {\mathcal{H}}_{\vartriangle }\) and \(\psi _{i}\in {\mathcal{H}}_{\triangledown }\). The derivative operators\(D^{\vartriangle }\) and \(D^{\triangledown }\) (the Malliavin derivatives) of functionals \(F_{t}\) and \(F_{\ast }\) of the above forms are defined as
and
respectively. These derivative operators \(D^{\vartriangle },D^{\triangledown }\) are then closable from \(L^{2}(\varOmega )\) into \(L^{2}(\varOmega; {\mathcal{H}}_{\vartriangle })\) and \(L^{2}(\varOmega;{\mathcal{H}}_{\ast })\), respectively. We denote by \({\mathbb{D}}^{\vartriangle ,1,2}\) and \({\mathbb{D}}^{\triangledown ,1,2}\) the closures of \({\mathcal{S}}_{\vartriangle }\) and \({\mathcal{S}}_{\triangledown }\) with respect to the norms
and
respectively. The divergence integrals\(\delta ^{\vartriangle }\) and \(\delta ^{\triangledown }\) are the adjoint of derivative operators \(D^{\vartriangle }\) and \(D^{\triangledown }\), respectively. We say that random variables \(v\in L^{2}(\varOmega;{\mathcal{H}}_{\vartriangle })\) and \(w\in L^{2}(\varOmega;{\mathcal{H}}_{\triangledown })\) belong to the domains of the divergence operators \(\delta ^{\vartriangle }\) and \(\delta ^{\triangledown }\), respectively, denoted by \({\mathrm{{Dom}}}(\delta ^{\vartriangle })\) and \({\mathrm{{Dom}}}(\delta ^{\triangledown })\) if
respectively, for all \(F_{\vartriangle }\in {\mathcal{S}}_{\vartriangle }\) and \(F_{\triangledown }\in {\mathcal{S}}_{\triangledown }\). In these cases \(\delta ^{\vartriangle }(v)\) and \(\delta ^{\triangledown }(w)\) are defined by the duality relationships
respectively, for any \(v\in {\mathbb{D}}^{\vartriangle ,1,2}\) and \(w\in {\mathbb{D}}^{\triangledown ,1,2}\). We have \({\mathbb{D}}^{\vartriangle ,1,2}\subset {\mathrm{{Dom}}}(\delta ^{\vartriangle })\) and \({\mathbb{D}}^{\triangledown ,1,2}\subset {\mathrm{{Dom}}}(\delta ^{\triangledown })\). We will use the notations
to express the Skorohod integrals, and the indefinite Skorohod integrals are defined as
respectively. It is important to note that we can localize the domains of the operators \(D^{\vartriangle },D^{\triangledown },\delta ^{\vartriangle }\) and \(\delta ^{\triangledown }\) via the usual manner. If \(\mathbb{L}\) is a class of random variables (or processes) we denote by \(\mathbb{L}_{\mathrm{loc}}\) the set of random variables F such that there exists a sequence \(\{(\varOmega _{n}, F^{n}), n\geq 1\}\subset {\mathscr{F}}\times \mathbb{L}\) with the following properties:
- (i)
\(\varOmega _{n}\uparrow \varOmega \),
- (ii)
\(F=F^{n}\) a.s. on \(\varOmega _{n}\).
Take the operators \(D^{\vartriangle }\) and \(\delta ^{\vartriangle }\) for instance. If \(F\in \mathbb{D}^{\vartriangle ,1,2}_{\mathrm{loc}}\), and \((\varOmega _{n}, F^{n})\) localizes F in \(\mathbb{D}^{\vartriangle , 1,2}\), then \(D^{\vartriangle }F\) is defined without ambiguity by \(D^{\vartriangle }F=D^{\vartriangle }F^{n}\) on \(\varOmega _{n}\), \(n\geq 1\). Then, if \(v\in {\mathbb{D}}^{\vartriangle ,1,2}_{\mathrm{loc}}\), the divergence \(\delta ^{\vartriangle }(v)\) is defined as a random variable determined by the conditions
where \((\varOmega _{n}, v^{n})\) is a localizing sequence for v, but it may depend on the localizing sequence.
At the end of this section, we introduce some estimates for the spatial process \(W=\{W_{x}=u(\cdot ,x)\}\) and temporal processes \(B=\{B_{t}=u(t,\cdot )\}\). For simplicity throughout this paper we let C stand for a positive constant depending only on the subscripts and its value may be different at different places, and this assumption also holds true for c. Moreover, the notation \(F\asymp G\) means that there are positive constants \(c_{1}\) and \(c_{2}\) such that
in the common domain of definition for F and G. We have
for all \(t\geq s>0\), \(x,y\in {\mathbb{R}}\) and
for all \(t>s>0\) and \(x,y\in {\mathbb{R}}\), where
and
for all \(t,s>0\) and \(z\geq 0\).
Lemma 2.1
For all \(t\geq s>0\) and \(x,y\in {\mathbb{R}}\) we have
By taking a special initial value (a two-sided \({\mathbb{R}}^{d}\)-valued Brownian motion), Mueller and Tribe [19] showed that the above estimate is reversible. However, when the initial value \(u(0,x)=0\), the above estimate is not reversible unless \(x=y\). In fact, we have
which implies that
with \(u=\frac{s}{t}\) by the continuity. On the other hand, we also have
for all \(t\geq s>0\) and \(u\in {\mathbb{R}}\). However, by the facts that
for all \(x\geq 0\), we have
for all \(t>0\) and \(z\geq 0\), and
for \(t>s\geq 0\) and \(z\geq 0\).
Corollary 2.1
For all \(t>0\) and \(x,y\in {\mathbb{R}}\) we have
Lemma 2.2
For all \(t,s,r>0\) and \(x\in {\mathbb{R}}\) we have
Proof
For all \(t,s,r>0\) and \(x\in {\mathbb{R}}\), we have
for all \(t,s,r>0\) and \(x\in {\mathbb{R}}\). □
Lemma 2.3
For all \(t>0\) and \(x,y,z\in {\mathbb{R}}\) we have
Proof
For all \(t>0\) and \(x,y,z\in {\mathbb{R}}\), we have
Consider the function \(f:{\mathbb{R}}_{+}\to {\mathbb{R}}_{+}\) defined by
Then, by the mean value theorem, we have
for all \(u,v\geq 0\) and some ξ between u and v. It follows that
for all \(t>0\) and \(x,y,z\in {\mathbb{R}}\). □
Lemma 2.4
For all \(t>s>t'>s'>0\) and \(x\in {\mathbb{R}}\) we have
Proof
For all \(t>s>t'>s'>0\) and \(x\in {\mathbb{R}}\) we have
Consider the function
with \(x\in [0,s]\). Then, by the mean value theorem, we have
for some \(s'\leq \xi \leq t'\) and the lemma follows. □
Lemma 2.5
For all \(t>0\) and \(x>y>x'>y'\) we have
Proof
Define the function \(f:{\mathbb{R}}_{+}\to {\mathbb{R}}_{+}\) by
Then, similar to the proof of Lemma 2.3, we have
for all \(t>0\) and \(x>y>x'>y'\). By mean value theorem it follows that
for some \(\xi \in [y,x]\). This completes the proof. □
Lemma 2.6
For all\(t>s>0\)and\(x\in {\mathbb{R}}\)denote\(\sigma ^{2}_{t,x}=E [u(t,x)^{2} ]\), \(\sigma ^{2}_{s,x}=E [u(s,x)^{2} ]\), \(\mu _{t,s,x}=E [u(t,x)u(s,x) ]\). Then we have
Proof
Given \(t>s>0\) and \(x\in {\mathbb{R}}\), we have
with \(z=\frac{s}{t}\). Elementary calculus can show that
for all \(0\leq z\leq 1\). In fact, we have
since \(\sqrt{z}+\sqrt{1-z^{2}}-1-\sqrt{z(1-z)}\geq 0\). Conversely, we have also
which gives
This completes the proof. □
Lemma 2.7
For all\(t>0\)and\(x>y\)denote\(\mu _{t,x,y}=E [u(t,x)u(t,y) ]\). Then, under the conditions of Lemma 2.6, we have
In particular, we have
for all\(t>0\)and\(x,y,z\in {\mathbb{R}}\).
Proof
Given \(t>0\) and \(x>y\). We have
by (2.5) and the lemma follows. □
3 The quadratic covariation of the spatial process \(W=\{W_{x},|x|\leq M\}\)
In this section, we study the existence of the PQC \([f(W),W]^{(SQ)}\). We fix a time parameter \(t>0\). Recall that
for \(\varepsilon >0\) and \(x\in {\mathbb{R}}\), and
provided the limit exists in probability. In [25], Pospisil and Tribe have showed that the following convergence holds.
Lemma 3.1
Let\(x\in {\mathbb{R}}\)and let\(x^{n}_{j}=\frac{jx}{n}; j=0,1,\ldots ,n\). Then we have
in\(L^{2}(\varOmega )\), asntends to infinity.
In this section we are not only to show that
in \(L^{2}\), but also to find a Banach space of Borel functions such that
exists in \(L^{2}\) for all Borel functions f belonging to the Banach space.
Proposition 3.1
We have
for all\(f\in C^{1}({\mathbb{R}})\).
Proof of Proposition 3.1
From Russo and Vallois [29] we only need to show that
almost surely. That is, for each \(t\geq 0\)
with some \(\alpha >0\), as \(\varepsilon \to 0\), by the above lemma, where
We have
for \(x\in {\mathbb{R}}\) and \(\varepsilon >0\), where
Recall that
where
Noting that
for all \(\varepsilon >0\) and \(y,z\in I_{x}\), we get
where \(\mu _{y,z}:=E [(W_{y+\varepsilon }-W_{y}) (W_{z+\varepsilon }-W_{z}) ]\).
Now, let us estimate the above function \(\varepsilon \mapsto \phi _{t,y}(\varepsilon )\). We have
which gives
Combining this with Lemma 2.5, we complete the proof. □
Now, we discuss the existence of the PQC \([f(W),W]^{(SQ)}\). Consider the decomposition
for \(\varepsilon >0\), and define the set
where
Then \({\mathscr{H}}_{t}=L^{2}({\mathbb{R}},\mu _{t}(dz))\) with
and \(\mu _{t}({\mathbb{R}})=C|x|<\infty \), which implies that the set \({\mathscr{E}}\) of elementary functions of the form
is dense in \({\mathscr{H}}_{t}\), where \(f_{i}\in {\mathbb{R}}\) and \(\{x_{i},0\leq i\leq l\}\) is a finite sequence of real numbers such that \(x_{i}< x_{i+1}\). Moreover, \({\mathscr{H}}_{t}\) includes all Borel functions f satisfying the condition
with \(0\leq \beta <\frac{\sqrt{\pi }}{4\sqrt{t}}\).
Theorem 3.1
Let\(f\in {\mathscr{H}}_{t}\). Then the PQC\([f(W), W]^{(SQ)}\)exists in\(L^{2}(\varOmega )\)and
for all\(x\in [-M,M]\).
In order to prove the theorem it is enough to prove the following two statements with \(f\in {\mathscr{H}}_{t}\):
- (1)
For any \(\varepsilon >0\) and \(x\in [-M,M]\), \(I_{\varepsilon }^{1,\pm }(f,x,\cdot )\in L^{2}(\varOmega )\). That is,
$$\begin{aligned} &E \bigl\vert I_{\varepsilon }^{1,-}(f,x,\cdot ) \bigr\vert ^{2}\leq CM \Vert f \Vert _{{ \mathscr{H}}_{t}}^{2}, \\ &E \bigl\vert I_{\varepsilon }^{1,+}(f,x,\cdot ) \bigr\vert ^{2}\leq CM \Vert f \Vert _{{ \mathscr{H}}_{t}}^{2}. \end{aligned}$$ - (2)
\(I_{\varepsilon }^{1,-}(f,x,t)\) and \(I_{\varepsilon }^{1,+}(f,x,t)\) are two Cauchy sequences in \(L^{2}(\varOmega )\) for all \(t>0\) and \(x\in [-M,M]\). That is,
$$ E \bigl\vert I_{\varepsilon _{1}}^{1,-}(f,x,t)-I_{\varepsilon _{2}}^{1,-}(f,x,t) \bigr\vert ^{2}\longrightarrow 0 $$and
$$ E \bigl\vert I_{\varepsilon _{1}}^{1,+}(f,x,t) -I_{\varepsilon _{2}}^{1,+}(f,x,t) \bigr\vert ^{2} \longrightarrow 0 $$for all \(x\in {\mathbb{R}}\), as \(\varepsilon _{1},\varepsilon _{2}\downarrow 0\).
We split the proof of the two statements into two parts.
Proof of statement (1)
Let \(x\in [-M,M]\). We have
for all \(\varepsilon >0\). Now, let us estimate the expression
for all \(\varepsilon _{1},\varepsilon _{2}>0\) and \(y,y'\in I_{x}\). To estimate the above expression, it is enough to assume that \(f\in {\mathscr{E}}\) by denseness, and moreover, by approximating we can assume that f is an infinitely differentiable function with compact support. It follows from the duality relationship (2.1) that
for all \(y,y'\in I_{x}\) and \(\varepsilon _{1},\varepsilon _{2}>0\). In order to end the proof we claim the
for all \(\varepsilon >0\) to be small enough.
For \(j=5\), from the fact
for \(0<|y-y'|\leq \varepsilon \), we have
for all \(\varepsilon >0\). Moreover, for \(|y-y'|>\varepsilon \) we have
by (2.8), which implies that
for all \(\varepsilon >0\). This shows that
for all \(\varepsilon >0\).
Next, let us estimate \(\sum_{j=1}^{4}\varLambda _{j}\). We have
and
where \(\rho ^{2}=\sigma ^{2}_{t,y}\sigma ^{2}_{t,y'}-\mu _{t,y,y'}^{2}\) and \(\varphi (x,y)\) is the density function of \(({W_{y}},{W_{y'}})\), that is,
Combining this with the identity
we get
A straightforward calculation shows that
for all \(m\geq 1\) and
since \(\sigma ^{2}_{t,y}=\sigma ^{2}_{t,y'}=\sqrt{\frac{t}{\pi }}\). It follows that
and
Thus, we get the estimate
and
by Lemma 2.7 and Lemma 2.3. Now, we can estimate \(\sum_{j=1}^{4}\varLambda _{j}\). We have
Combining this with (3.9), (3.10), Lemma 2.3 and Lemma 2.5, we get
for all \(\varepsilon >0\) and \(x\in {\mathbb{R}}\). This shows that
Similarly, one can show the estimate
and the first statement follows. □
Proof of statement (2)
Given \(x\in [-M,M]\). Without loss of generality we assume that \(\varepsilon _{1}>\varepsilon _{2}\). We prove only the first convergence and similarly one can prove the second convergence. We have
for all \(\varepsilon _{1},\varepsilon _{2}>0\), where
and
To end the proof, it is enough to assume that \(f\in {\mathscr{E}}\) by denseness, and moreover, by approximating we can assume that f is an infinitely differentiable function with compact support. It follows from (3.8) that
which gives
with \(i,j\in \{1,2\}\) and \(i\neq j\), where
for all \(\varepsilon _{1},\varepsilon _{2}>0\) and \(y,y'\in {\mathbb{R}}\). Now, we claim that the following convergences hold:
with \(i,j\in \{1,2\}\) and \(i\neq j\), as \(\varepsilon _{1},\varepsilon _{2}\to 0\). We decompose
with \(i,j\in \{1,2\}\) and \(i\neq j\). By symmetry we only need to prove the convergence (3.11) with \(i=1\) and \(j=2\).
Step I. The following convergence holds:
as \(\varepsilon _{1},\varepsilon _{2}\to 0\). We have
Consider the next function on \({\mathbb{R}}_{+}\) (see Sect. 2):
Then we have
Notice that, by Taylor’s expansion,
One gets
with \(\alpha _{1}=\frac{1}{4t^{2}} o (12(y-y')^{2}\varepsilon _{1}^{2}+2 \varepsilon _{1}^{4} )\) and
with \(\alpha _{2}=\frac{1}{4t^{2}} o (12(y-y')^{2}\varepsilon _{1} \varepsilon _{2} +\varepsilon _{1}\varepsilon _{2}(\varepsilon _{1}- \varepsilon _{2}) (12(y-y')+ 2\varepsilon _{1}-\varepsilon _{2} ) )\). It follows that
for all \(0<\varepsilon _{2}<\varepsilon _{1}<1\) and \(y,y'\in I_{x}\), which shows that the convergence (3.11) holds since \(f\in {\mathscr{H}}_{t}\).
Step II. The following convergence holds:
as \(\varepsilon _{1},\varepsilon _{2}\to 0\). Keeping the notations in Step I, we have
for all ε and
It follows from (2.8) that
for all \(0<\varepsilon _{2}<\varepsilon _{1}<1\) and \(y,y'\in I_{x}\), which implies that
as \(\varepsilon _{1},\varepsilon _{2}\to 0\) since \(f\in {\mathscr{H}}_{t}\). Similarly, we can show the next convergence:
as \(\varepsilon _{1},\varepsilon _{2}\to 0\).
Step III. The following convergence holds:
as \(\varepsilon _{1},\varepsilon _{2}\to 0\). By Step II and Lemma 2.3, we have
for all \(0<\varepsilon _{2}<\varepsilon _{1}<1\) and \(y,y'\in I_{x}\), which implies that
as \(\varepsilon _{1},\varepsilon _{2}\to 0\) by (3.9), since \(f\in {\mathscr{H}}_{t}\). Similarly, we can show that the next convergence:
as \(\varepsilon _{1},\varepsilon _{2}\to 0\). Thus, we have proved the second statement. □
Corollary 3.1
Let\(f,f_{1},f_{2},\ldots \in {\mathscr{H}}_{t}\)such that\(f_{n}\to f\)in\({\mathscr{H}}_{t}\). Then the convergence
holds in\(L^{2}(\varOmega )\)for all\(x\in {\mathbb{R}}\).
4 The Itô formula for the spatial process
In this section, as an application of the previous section we discuss the Itô calculus for the process \(W=\{W_{x},|x|\leq M\}\) and fix a time parameter \(t>0\). For a continuous process X admitting a finite quadratic variation \([X,X]\), Russo and Vallois [29] have introduced the following Itô formula:
for all \(F\in C^{2}({\mathbb{R}})\), where
is called the forward integral, where the notation ucplim denotes the uniform convergence in probability on each compact interval. We refer to Russo and Vallois [29] and the references therein for more details of stochastic calculus of continuous processes with finite quadratic variations. It follows from the previous section (the quadratic variation of \(\{W_{x},x\in {\mathbb{R}}\}\) is equal to \(|x|\) for all \(x\in {\mathbb{R}}\)) that
for all \(F\in C^{2}({\mathbb{R}})\). Thus, by smooth approximating, we have the next Itô type formula.
Theorem 4.1
Let\(f\in {\mathscr{H}}_{t}\)be left continuous. IfFis an absolutely continuous function with the derivative\(F'=f\), then the following Itô type formula holds:
Clearly, this is an analogue of Föllmer–Protter–Shiryayev’s formula. It is an improvement in terms of the hypothesis on f and it is also quite interesting itself. Some details and more work could be found in Eisenbaum [6], Feng–Zhao [8], Föllmer et al. [9], Moret–Nualart [18], Peskir [24], Rogers–Walsh [26], Russo–Vallois [28, 29], Yan et al. [36, 37], and the references therein. It is well known that when W is a semimartingale, the forward integral coincides with the Itô integral. However, the following theorem points out that the two integrals are coincident for the process \(W=\{W_{x},x\in {\mathbb{R}}\}\). But \(W=\{W_{x},x\in {\mathbb{R}}\}\) is not a semimartingale.
Proof of Theorem 4.1
If \(f\in C^{1}({\mathbb{R}})\), then this is Itô’s formula since
For \(f\notin C^{1}({\mathbb{R}})\), by a localization argument we may assume that the function f is uniformly bounded. In fact, for any \(k\geq 0\) we may consider the set
and let \(f^{[k]}\) be a measurable function such that \(f^{[k]}=f\) on \([-k,k]\) and such that \(f^{[k]}\) vanishes outside. Then \(f^{[k]}\) is uniformly bounded and \(f^{[k]}\in {\mathscr{H}}_{t}\) for every \(k\geq 0\). Set \(\frac{d}{dx}F^{[k]}=f^{[k]}\) and \(F^{[k]}=F\) on \([-k,k]\). If the theorem is true for all uniformly bounded functions on \({\mathscr{H}}_{t}\), then we get the desired formula
on the set \(\varOmega _{k}\). Letting k tend to infinity we deduce the Itô formula (4.1).
Let now \(F'=f\in {\mathscr{H}}_{t}\) be uniformly bounded and left continuous. Consider the function ζ on \({\mathbb{R}}\) by
where c is a normalizing constant such that \(\int _{\mathbb{R}}\zeta (x)\,dx=1\). Define the mollifiers
and the sequence of smooth functions
Then \(F_{n}\in C^{\infty }({\mathbb{R}})\) for all \(n\geq 1\) and the Itô formula
holds for all \(n\geq 1\), where \(f_{n}=F_{n}'\). Moreover, by using Lebesgue’s dominated convergence theorem, one can prove that, as \(n\to \infty \), for each x,
and \(\{f_{n}\}\subset {\mathscr{H}}_{t}\), \(f_{n}\to f\) in \({\mathscr{H}}_{t}\). It follows that
and
in \(L^{2}(\varOmega )\) by Corollary 3.1, as n tends to infinity. It follows that
in \(L^{2}(\varOmega )\), as n tends to infinity. This completes the proof since the integral is closed in \(L^{2}(\varOmega )\). □
Now, we consider the Itô formula including the Skorohod integral of the spatial process \(W=\{W_{x}\}\).
Theorem 4.2
Letfbe left continuous. IfFis an absolutely continuous function with the derivative\(F'=f\)satisfying the condition
with\(0\leq \beta <\frac{\sqrt{\pi }}{4\sqrt{t}}\), then the following Itô type formula holds:
for all\(x\in [-M,M]\).
According to the two theorems above we get the next relationship:
if f satisfies the growth condition (4.6). Similar to the proof of Theorem 4.1 one can introduce Theorem 4.2. But we need to give the following standard Itô type formula:
for all \(F\in C^{2}({\mathbb{R}})\) satisfying the condition
with \(0\leq \beta <\frac{\sqrt{\pi }}{4\sqrt{t}}\). It is important to note that one have given a standard Itô formula for a large class of Gaussian processes in Alós et al. [1]. However, the process \(x\mapsto W_{x}\) does not satisfy the condition in Alós et al. [1] since
for all \(t\geq 0\) and \(x\in {\mathbb{R}}\). So, we need to give the proof of the formula (4.9).
Proof of (4.9)
Let us fix \(x\in [-M,M]\) and let \(\pi \equiv \{x^{n}_{j}=\frac{jx}{n}; j=0,1,\ldots ,n\}\) be a partition of \([0,x]\). Clearly, the growth condition (4.6) implies that
for some constant \(c>0\) and all \(p<\frac{\sqrt{\pi }}{2\beta \sqrt{t}}\), where \(G\in \{F,F',F''\}\). In particular, the estimate (4.11) holds for \(p=2\). Using the Taylor expansion, we have
where \(W_{j}(\theta _{j})=W_{x^{n}_{j-1}}+\theta _{j}(W_{x^{n}_{j}}-W_{x^{n}_{j-1}})\) with \(\theta _{j}\in (0,1)\) being a random variable. By the duality relationship (2.1) we have
Now, in order to end the proof we claim that the following convergences in \(L^{2}\) hold:
as n tends to infinity.
To prove the first convergence, it is enough to establish that
as n tends to infinity. By the -Minkowski inequality we have
Now, we prove the third convergence. We have
Suppose that \(n\geq m\), and for any \(j=1,\ldots ,n\) we denote by \(x^{m(n)}_{j}\) the point of the mth partition that is closer to \(x^{n}_{j}\) from the left hand. Then we obtain
Clearly, we have \(\varLambda _{n}(2,2)\to 0\) (\(n,m\to \infty \)) by Lemma 3.1 and the estimate (4.11),
and
by (2.4), the estimate (4.11) and Minkowski’s inequality. Thus, we obtain the third convergence, i.e., \(J^{n}\to \int ^{t}_{0}F^{\prime \prime }(W_{y})\,dy\) in \(L^{1}\).
Finally, to end the proof we address the second convergence:
We need to show that
in \(L^{2}(\varOmega;{\mathcal{H}}_{\vartriangle })\), as n tends to infinity. We have
for all \(n\geq 1\) and \(x\in {\mathbb{R}}\). Notice that
by (2.6) and (2.8). Combining this with (4.16) and the estimate (4.11), we get
in \(L^{2}(\varOmega;{\mathcal{H}}_{\vartriangle })\), as n tends to infinity. It follows that
in \(L^{2}(\varOmega )\), as n tends to infinity. This completes the proof since the integral \(\int _{0}^{\cdot }u_{s}\delta W_{s}\) is closed in \(L^{2}(\varOmega )\). □
5 The Bouleau–Yor identity for the spatial process
In this section, we consider the local time of the process \(W=\{W_{x},x\in [0,M]\}\). Our main object is to prove that the integral
is well-defined and that the identity
holds for all \(g\in {\mathscr{H}}_{t}\) and \(t>0\), where
is the local time of \(W=\{W_{x},x\in [0,M]\}\). The identity (5.1) is called the Bouleau–Yor identity. More work on this subject can be found in Bouleau–Yor [3], Eisenbaum [6], Föllmer et al. [9], Feng–Zhao [8], Peskir [24], Rogers–Walsh [26], Yan et al. [36, 37], and the references therein.
Recall that, for any closed interval \(I\subset {\mathbb{R}}_{+}\) and for any \(a\in {\mathbb{R}}\), the local time \(L(a,I)\) of u is defined as the density of the occupation measure \(\mu _{I}\) defined by
It can be shown (see Geman and Horowitz [10], Theorem 6.4) that the following occupation density formula holds:
for every Borel function \(g(a,x)\geq 0\) on \(I\times {\mathbb{R}}\). Thus, some estimates in Sect. 2 and Theorem 21.9 in Geman–Horowitz [10] together imply that the following result holds.
Corollary 5.1
The local time\({\mathscr{L}}^{\vartriangle }(a,x):=L(a, [0,x])\)of\(W=\{W_{x},|x|\leq M\}\)exists and\({\mathscr{L}}^{t}\in L^{2}(\lambda \times P)\)for all\(x\in [0,M]\)and\((a,x)\mapsto {\mathscr{L}}^{\vartriangle }(a,x)\)is jointly continuous, whereλdenotes Lebesgue measure. Moreover, the occupation formula
holds for every continuous and bounded function\(\psi (a,x):{\mathbb{R}}\times {\mathbb{R}}_{+}\rightarrow {\mathbb{R}}\)and any\(x\in [0,M]\).
Lemma 5.1
For any\(f_{\triangle }=\sum_{j}f_{j}1_{(a_{j-1},a_{j}]}\in {\mathscr{E}}\), we define
Then the integral is well-defined and
almost surely, for all\(x\in [0,M]\).
Proof
For the function \(f_{\triangle }(y)=1_{(a,b]}(y)\) we define the sequence of smooth functions \(f_{n}, n=1,2,\ldots \) , by
for all \(y\in \mathbb{R}\), where \(\zeta _{n},n\geq 1\) are the so-called mollifiers given in (4.4). Then \(\{f_{n}\}\subset C^{\infty }({\mathbb{R}})\cap {\mathscr{H}}_{t}\) and \(f_{n}\) converges to \(f_{\triangle }\) in \({\mathscr{H}}_{t}\), as n tends to infinity. It follows from the occupation formula that
almost surely, as \(n\to \infty \), by the continuity of \(y\mapsto {\mathscr{L}}^{\vartriangle }(y,x)\). On the other hand, we see also that there exists a subsequence \(\{f_{n_{k}}\}\) such that
for all \(x\geq 0\), almost surely, as \(k\to \infty \) since \(f_{n}\) converges to \(f_{\triangle }\) in \({\mathscr{H}}_{t}\). It follows that
for all \(x\geq 0\), almost surely. Thus, the identity
follows from the linearity property, and the lemma follows. □
As a direct consequence of the above lemma, for every \(f\in {\mathscr{H}}_{t}\) if
in \({\mathscr{H}}_{t}\), where \(\{f_{\triangle ,n}\},\{g_{\triangle ,n}\}\subset {\mathscr{E}}\), we then have
in \(L^{2}(\varOmega )\). Thus, by the denseness of \({\mathscr{E}}\) in \({\mathscr{H}}_{t}\) we can define
for any \(f\in {\mathscr{H}}_{t}\), where \(\{f_{\triangle ,n}\}\subset {\mathscr{E}}\) and
in \({\mathscr{H}}_{t}\). The considerations are enough to prove the following theorem.
Theorem 5.1
For any\(f\in {\mathscr{H}}_{t}\), the integral
is well-defined in\(L^{2}(\varOmega )\)and the Bouleau–Yor type identity
holds, almost surely, for all\(x\in [0,M]\).
Corollary 5.2
(Tanaka formula)
For any \(a\in {\mathbb{R}}\) we have
Proof
Take \(F(y)=(y-x)^{+}\). Then F is absolutely continuous and
It follows from the identity (5.3) and Itô’s formula (4.2) that
for all \(x\geq 0\), which gives the first identity. In the same way one can obtain the second identity, and by subtracting the last identity from the previous one, we get the third identity. □
According to Theorem 5.1, we get an analogue of the Itô formula (Bouleau–Yor type formula).
Corollary 5.3
Let\(f\in {\mathscr{H}}_{t}\)be a left continuous function with right limits. IfFis an absolutely continuous function with\(F'=f\), then the Itô type formula
holds for all\(x\geq 0\).
Recall that, if F is the difference of two convex functions, then F is an absolutely continuous function with derivative of bounded variation. Thus, the Itô–Tanaka formula
holds.
6 The quadratic covariation of process \(B=\{B_{t},t\geq 0\}\)
In this section, we study the existence of the PQC \([f(B),B]^{(TQ)}\). Recall that
for \(\varepsilon >0,t\geq 0\) and \(x\in {\mathbb{R}}\), and
provided the limit exists in probability. In this section, we study some analysis questions of the process \(\{u(t,\cdot ),t\geq 0\}\) associated with the quadratic covariation \([f(u(\cdot ,x)),u(\cdot ,x)]^{(TQ)}\), and the research includes the existence of the PQC \([f(u(\cdot ,x)),u(\cdot ,x)]^{(TQ)}\), the Itô and Tanaka formulas. Recall that \(B_{t}=u(t,\cdot )\)\((t\in [0,T])\) and
for all \(t\geq 0\) and \(x\in {\mathbb{R}}\). It follows from Alós et al. [1] that the Itô formula holds,
for all \(t\in [0,T]\) and \(f\in C^{2}({\mathbb{R}})\) satisfying the condition
with \(0\leq \beta <\frac{\sqrt{\pi }}{4\sqrt{T}}\).
Recall that the local Hölder index \(\gamma _{0}\) of a continuous paths process \(\{X_{t}: t\geq 0\}\) is the supremum of the exponents γ verifying, for any \(T>0\),
Lemma 6.1
(Gradinaru–Nourdin [11])
Let\(g:{\mathbb{R}}\to {\mathbb{R}}\)be a function satisfying
for all\(x,y\in {\mathbb{R}}\)and letXbe a locally Hölder continuous paths process with index\(\gamma \in (0,1)\). Assume thatVis a bounded variation continuous paths process. Set
for\(t\geq 0\), \(\varepsilon >0\). If for each\(t\geq 0\), as\(\varepsilon \to 0\),
with\(\alpha >0\), then\(\lim_{\varepsilon \to 0}X^{g}_{\varepsilon }(t)=V_{t}\)almost surely, for any\(t\geq 0\), and ifgis non-negative, for any continuous stochastic process\(\{Y_{t}: t\geq 0\}\),
almost surely, uniformly inton each compact interval.
Proposition 6.1
Let\(f\in C^{1}({\mathbb{R}})\). We have
and, in particular, we have
for all\(t\geq 0\).
Proof
By Lemma 6.1 and the Hölder continuity of the solution \(u(t,x)\), it is enough to show that the estimate
holds, for each \(t\geq 0\), with some \(\alpha >0\), as \(\varepsilon \to 0\), where
We have
for \(t\geq 0\) and \(\varepsilon >0\), where
Define the function \(\phi _{s}: {\mathbb{R}}_{+}\to {\mathbb{R}}_{+}\) by
for every \(s>0\). Then we have
for all \(s>0\). Noting that
for all \(r,s\geq 0\) and \(\varepsilon >0\), we get
where \(\mu _{s,r}:=E [(B_{s+\varepsilon }-B_{s}) (B_{r+\varepsilon }-B_{r}) ]\). Now, let us estimate the function
Clearly, one can see that
and the continuity of the function \(x\mapsto 1-2\sqrt{1-x/2}+\sqrt{1-x}\) implies that
with \(x=\frac{\varepsilon }{s+\varepsilon }\) and \(0<\beta <\frac{1}{2}\), which gives
It follows from Lemma 2.4 that there is a constant \(\alpha >0\) such that
for all \(t>0\), which gives the desired estimate
for each \(t\geq 0\) and some \(\alpha >0\). This completes the proof. □
Consider the decomposition
for \(\varepsilon >0\), and by estimating the two terms in the right hand side above in \(L^{2}(\varOmega )\) one can structure the next Banach space:
where
Clearly, \({\mathscr{H}}_{\ast }=L^{2}({\mathbb{R}},\mu (dz))\) with
and \({\mathscr{H}}_{\ast }\) includes all functions f satisfying the condition
with \(0\leq \beta <\frac{\sqrt{\pi }}{4\sqrt{T}}\). In the same way as proving Theorem 3.1 and by smooth approximation one can introduce the following result.
Theorem 6.1
The PQC\([f(B),B]^{(TQ)}\)exists and
for all\(f\in {\mathscr{H}}_{\ast }\)and\(t\in [0,T]\). Moreover, ifFis an absolutely continuous function such that
with\(0\leq \beta <\frac{\sqrt{\pi }}{4\sqrt{T}}\), then the following Itô type formula holds:
for all\(t\in [0,T]\).
Recall that Russo and Tudor [27] have showed that \(B=\{B_{t}=u(t,\cdot ),t\geq 0\}\) admits a local time \(L(t,a)\in L^{2}(\lambda \times P)\) such that \((a,t)\mapsto L(a,t)\) is jointly continuous, where λ denotes the Lebesgue measure, since \(B=\{B_{t}=u(t,\cdot ),t\geq 0\}\) is a bi-fractional Brownian motion for every \(x\in {\mathbb{R}}\). Define the weighted local time \({\mathscr{L}}\) of \(B=\{B_{t}=u(t,\cdot ),t\geq 0\}\) by
for \(t\geq 0\) and \(x\in {\mathbb{R}}\), where δ is the Dirac delta function. Then the occupation formula
holds for every continuous and bounded function \(\psi:{\mathbb{R}}\times {\mathbb{R}}_{+}\rightarrow {\mathbb{R}}\) and any \(x\geq 0\). As in Sect. 5, we can show that the integral
is well-defined and
almost surely, for all \(f_{\triangle }=\sum_{j}f_{j}1_{(a_{j-1},a_{j}]}\in {\mathscr{E}}\). By the denseness of \({\mathscr{E}}\) in \({\mathscr{H}}_{\ast }\) one can define
for any \(f\in {\mathscr{H}}_{\ast }\), where \(\{f_{\triangle ,n}\}\subset {\mathscr{E}}\) and
in \({\mathscr{H}}\). Moreover, the Bouleau–Yor type formula
holds, almost surely, for all \(f\in {\mathscr{H}}_{\ast }\).
Corollary 6.1
(Tanaka formula)
For any \(x\in {\mathbb{R}}\) we have
References
Alós, E., Mazet, O., Nualart, D.: Stochastic calculus with respect to Gaussian processes. Ann. Probab. 29, 766–801 (2001)
Balan, R., Kim, D.: The stochastic heat equation driven by a Gaussian noise: germ Markov property. Commun. Stoch. Anal. 2, 229–249 (2008)
Bouleau, N., Yor, M.: Sur la variation quadratique des temps locaux de certaines semimartingales. C. R. Acad. Sci., Sér. 1 Math. 292, 491–494 (1981)
Da Prato, G., Jentzen, A., Röckner, M.: A mild Itô formula for SPDEs (2012). arXiv:1009.3526
Deya, A., Tindel, S.: Malliavin Calculus for Fractional Heat Equation, Malliavin Calculus and Stochastic Analysis, 361-384. Springer Proc. Math. Stat., vol. 34. Springer, New York (2013)
Eisenbaum, N.: Integration with respect to local time. Potential Anal. 13, 303–328 (2000)
Errami, M., Russo, F.: n-covariation, generalized Dirichlet processes and calculus with respect to finite cubic variation process. Stoch. Process. Appl. 104, 259–299 (2003)
Feng, C.R., Zhao, H.Z.: Two-parameters \(p,q\)-variation paths and integrations of local times. Potential Anal. 25, 165–204 (2006)
Föllmer, H., Protter, P., Shiryayev, A.N.: Quadratic covariation and an extension of Itô’s formula. Bernoulli 1, 149–169 (1995)
Geman, D., Horowitz, J.: Occupation densities. Ann. Probab. 8, 1–67 (1980)
Gradinaru, M., Nourdin, I.: Approximation at first and second order of m-order integrals of the fractional Brownian motion and of certain semimartingales. Electron. J. Probab. 8, 1–26 (2003)
Gradinaru, M., Nourdin, I., Tindel, S.: Ito’s- and Tanaka’s-type formulae for the stochastic heat equation: the linear case. J. Funct. Anal. 228, 114–143 (2005)
Houdré, C., Villa, J.: An example of infinite dimensional quasi-helix. In: Stochastic Models, Mexico City, 2002. Contemp. Math., vol. 336, pp. 195–201. Amer. Math. Soc., Providence (2003)
Lanconelli, A.: White noise approach to the Itô formula for the stochastic heat equation. Commun. Stoch. Anal. 1, 311–320 (2007)
Lanconelli, A.: A remark on the renormalized square of the solution of the stochastic heat equation and on its associated evolution. Infin. Dimens. Anal. Quantum Probab. Relat. Top. 12, 497–502 (2009)
Lei, P., Nualart, D.: A decomposition of the bi-fractional Brownian motion and some applications. Stat. Probab. Lett. 79, 619–624 (2009)
León, J.A., Tindel, S.: Itô formula for linear fractional PDEs. Stochastics 80, 427–450 (2008)
Moret, S., Nualart, D.: Quadratic covariation and Itô’s formula for smooth nondegenerate martingales. J. Theor. Probab. 13, 193–224 (2000)
Mueller, C., Tribe, R.: Hitting probabilities of a random string. Electron. J. Probab. 7, 10 (2002)
Nualart, D.: Malliavin Calculus and Related Topics, 2nd edn. Springer, New York (2006)
Nualart, D., Vuillermot, P.-A.: Variational solutions for partial differential equations driven by a fractional noise. J. Funct. Anal. 232, 390–454 (2006)
Ouahhabi, H., Tudor, C.A.: Additive functionals of the solution to fractional stochastic heat equation. J. Fourier Anal. Appl. 19, 777–791 (2013)
Pardoux, E.: Two-sided stochastic calculus for SPDEs. In: Stochastic Partial Differential Equations and Applications, Trento, 1985. Lecture Notes in Math., vol. 1236, pp. 200–207. Springer, Berlin (1987)
Peskir, G.: A change-of-variable formula with local time on curves. J. Theor. Probab. 18, 499–535 (2005)
Pospisil, J., Tribe, R.: Parameter estimation and exact variations for stochastic heat equations driven by space-time white noise. Stoch. Anal. Appl. 4, 830–856 (2007)
Rogers, C.G., Walsh, J.B.: Local time and stochastic area integrals. Ann. Probab. 19, 457–482 (1991)
Russo, F., Tudor, C.A.: On bifractional Brownian motion. Stoch. Process. Appl. 116, 830–856 (2006)
Russo, F., Vallois, P.: Itô formula for \({\mathcal{C}}^{1}\)-functions of semimartingales. Probab. Theory Relat. Fields 104, 27–41 (1996)
Russo, F., Vallois, P.: Stochastic calculus with respect to a continuous finite quadratic variation process. Stoch. Stoch. Rep. 70, 1–40 (2000)
Swanson, J.: Variations of the solution to a stochastic heat equation. Ann. Probab. 35, 2122–2159 (2007)
Torres, S., Tudor, C.A., Viens, F.G.: Quadratic variations for the fractional-colored stochastic heat equation. Electron. J. Probab. 19(76), 1–51 (2014)
Tudor, C.A.: Chaos expansion and regularity of the local time of the solution to the stochastic heat equation with additive fractional-colored noise. Taiwan. J. Math. 17, 1765–1777 (2013)
Tudor, C.A., Xiao, Y.: Some path properties of bi-fractional Brownian motion. Bernoulli 13, 1023–1052 (2007)
Tudor, C.A., Xiao, Y.: Sample paths of the solution to the fractional-colored stochastic heat equation. Stoch. Dyn. 17(03), 1750004 (2017)
Yan, L., Gao, B., Liu, J.: The Bouleau–Yor identity for a bi-fBm. Stochastics 86, 382–414 (2014)
Yan, L., Liu, J., Chen, C.: The generalized quadratic covariation for fractional Brownian motion with Hurst index less than \(1/2\). Infin. Dimens. Anal. Quantum Probab. Relat. Top. 17(4), 32 (2014)
Yan, L., Liu, J., Yang, X.: Integration with respect to fractional local time with Hurst index \(1/2< H<1\). Potential Anal. 30, 115–138 (2009)
Zambotti, L.: Itô–Tanaka’s formula for stochastic partial differential equations driven by additive space-time white noise. In: SPDEs and Applications VII. Lect. Notes Pure Appl. Math., vol. 245, pp. 337–347. Chapman & Hall, Boca Raton (2006)
Acknowledgements
The authors are thankful to anonymous referees for their careful corrections, helpful suggestions, and valuable comments on the original version of this paper.
Availability of data and materials
Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.
Funding
The project was sponsored by NSFC (Nos. 11971101, 11571071, 11626033, 11701589), Innovation Program of the Shanghai Municipal Education Commission (12ZZ063), Natural Science Foundation of Anhui Province (1808085MA02) and Quality engineering of Anhui Education (2018jyssf045).
Author information
Authors and Affiliations
Contributions
All authors contributed equally to the manuscript and typed, read, and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Sun, X., Yan, L. & Yu, X. Quadratic covariations for the solution to a stochastic heat equation with space-time white noise. Adv Differ Equ 2020, 254 (2020). https://doi.org/10.1186/s13662-020-02707-9
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s13662-020-02707-9