In this paper, we focus on a class of doubly perturbed stochastic differential equations with Markovian switching. Using the Girsanov transformation argument we establish the quadratic transportation inequalities for the law of the solution of those equations with Markovian switching under the d2 metric and the uniform metric d∞.
Citation: Liping Xu, Zhi Li, Weiguo Liu, Jie Zhou. Transportation inequalities for doubly perturbed stochastic differential equations with Markovian switching[J]. AIMS Mathematics, 2021, 6(3): 2874-2885. doi: 10.3934/math.2021173
[1] | Zhengqi Zhang, Huaiqin Wu . Cluster synchronization in finite/fixed time for semi-Markovian switching T-S fuzzy complex dynamical networks with discontinuous dynamic nodes. AIMS Mathematics, 2022, 7(7): 11942-11971. doi: 10.3934/math.2022666 |
[2] | Yuxiao Zhao, Hong Lin, Xiaoyan Qiao . Persistence, extinction and practical exponential stability of impulsive stochastic competition models with varying delays. AIMS Mathematics, 2023, 8(10): 22643-22661. doi: 10.3934/math.20231152 |
[3] | Hui Sun, Zhongyang Sun, Ya Huang . Equilibrium investment and risk control for an insurer with non-Markovian regime-switching and no-shorting constraints. AIMS Mathematics, 2020, 5(6): 6996-7013. doi: 10.3934/math.2020449 |
[4] | Li Liu, Yinfang Song, Hong Yu, Gang Zhang . Almost sure exponential synchronization of multilayer complex networks with Markovian switching via aperiodically intermittent discrete observa- tion noise. AIMS Mathematics, 2024, 9(10): 28828-28849. doi: 10.3934/math.20241399 |
[5] | Qiang Li, Weiqiang Gong, Linzhong Zhang, Kai Wang . Robust dissipativity and passivity of stochastic Markovian switching CVNNs with partly unknown transition rates and probabilistic time-varying delay. AIMS Mathematics, 2022, 7(10): 19458-19480. doi: 10.3934/math.20221068 |
[6] | Jingjing Yang, Jianqiu Lu . Stabilization in distribution of hybrid stochastic differential delay equations with Lévy noise by discrete-time state feedback controls. AIMS Mathematics, 2025, 10(2): 3457-3483. doi: 10.3934/math.2025160 |
[7] | Xipan Zhang, Changchun Shen, Dingju Xu . Reachable set estimation for neutral semi-Markovian jump systems with time-varying delay. AIMS Mathematics, 2024, 9(4): 8043-8062. doi: 10.3934/math.2024391 |
[8] | Nuraddeen S. Gafai, Ali H. M. Murid, Samir Naqos, Nur H. A. A. Wahid . Computing the zeros of the Szegö kernel for doubly connected regions using conformal mapping. AIMS Mathematics, 2023, 8(5): 12040-12061. doi: 10.3934/math.2023607 |
[9] | Yuanfu Shao . Dynamics and optimal harvesting of a stochastic predator-prey system with regime switching, S-type distributed time delays and Lévy jumps. AIMS Mathematics, 2022, 7(3): 4068-4093. doi: 10.3934/math.2022225 |
[10] | Yakufu Kasimu, Gulijiamali Maimaitiaili . Non-fragile H∞ filter design for uncertain neutral Markovian jump systems with time-varying delays. AIMS Mathematics, 2024, 9(6): 15559-15583. doi: 10.3934/math.2024752 |
In this paper, we focus on a class of doubly perturbed stochastic differential equations with Markovian switching. Using the Girsanov transformation argument we establish the quadratic transportation inequalities for the law of the solution of those equations with Markovian switching under the d2 metric and the uniform metric d∞.
Consider the following doubly perturbed stochastic differential equations with Markovian switching
X(t)=x+∫t0f(s,X(s),r(s))ds+∫t0g(s,X(s),r(s))dB(s)+αmax0≤s≤tX(s)+βmin0≤s≤tX(s). | (1.1) |
where r(t) be a right continuous Markov chain taking values in a finite state space S={1,2,⋯,N}, f:[0,T]×R×S→R and g:[0,T]×R×S→R are some appropriate functions; W is a standard Brownian motion and α and β are two constants. There now exists a considerable body of literatures which are devoted to the study of "perturbed" stochastic equations. As the limit process from a weak polymers model, Carmona et al. [5] and Norris et al. [19] investigated the following doubly perturbed Brownian motion
x(t)=B(t)+αmax0≤s≤tx(s)+βmin0≤s≤tx(s). | (1.2) |
Many researchers have devoted themselves to studying this model due to its extensive applications, see Davis [7], Carmona, Petit and Yor [4], Perman and Werner [23], Chaumont and Doney [6], etc. Following them, Doney and Zhang [8] studied the following singly perturbed Skorohod equations
x(t)=x0+∫t0g(s,x(s))dB(s)+∫t0f(s,x(s))ds+αmax0≤s≤tx(s). | (1.3) |
The authors proved the existence and uniqueness of the solution for (1.3) when the coefficients b, σ are the global Lipschitz. Mao et al. [18] discussed the following doubly perturbed stochastic differential equations
x(t)=x0+∫t0g(s,x(s))dB(s)+∫t0f(s,x(s))ds+αmax0≤s≤tx(s)+βmin0≤s≤tx(s). | (1.4) |
They showed the existence and uniqueness of the solution for (1.4) under some non-Lipschitz conditions. Hu and Ren [13] and Luo [16] extended the equation (1.4) to the neutral type and doubly perturbed jump-diffusion processes, respectively. Liu and Yang [15] established the existence and uniqueness of the solution for a class of doubly perturbed neutral jump-diffusion processes with Markovian switching.
On the other hand, in the past decade, a plenty of results have been published concerning Talagrand-type transportation cost inequalities (TCIs) on the path spaces of stochastic processes. Now let us consider the kinds of inequalities we will deal with. Let (E,d) be a metric space equipped with a σ-field B such that the distance d(⋅,⋅) is B⊗B-measurable. Given p≥1 and two probability measures μ and ν on E, the Wasserstein distance is defined by
Wdp(μ,ν)=infπ∈C(μ,ν)(∫∫d(x,y)pdπ(x,y))1/p, |
where C(μ,ν) denotes the totality of probability measures on the product space E×E with the marginals μ and ν. The relative entropy of ν with respect to μ is defined as
H(ν|μ)={∫lndνdμdν,ν≪μ,+∞,otherwise. |
The probability measure μ satisfies the Tp transportation inequality on (E,d) if there exists a constant C≥0 such that for any probability measure ν,
Wdp(μ,ν)≤√2CH(ν|μ). |
As usual, we write μ∈Tp(C) for this relation. As is known to all, the cases "p=1" and "p=2" are of particular interest. T1(C) is related to concentration of measure phenomenon and well characterized, as it was shown by Djellout [9] using preliminary results obtained in Bobkov and Götze [1].
The property T2(C) is stronger than T1(C) and it has the dimension free tensorization property. The property T2(C) is closely linked with many other functional properties such as Poincaré inequality, logarithmic Sobolev inequality and Hamilton-Jacobi equations. For example, Otto and Villani [20] showed that in a smooth Riemannian setting, the logarithmic Sobolev inequality implies T2(C), and T2(C) means the Poincaré inequality. Since Talagrand [25] established T2(C) for the Gaussian measure which was generalized in [11] to the framework of an abstract Wiener space, in the past decade, a plenty of results have been published concerning T2(C) on the path spaces of stochastic processes, see e.g. [9,30,31] for diffusion processes on Rd, [21] for multidimensional semi-martingales, [26] for diffusion processes with history-dependent drift, [27,28] for diffusion processes on Riemannian manifolds, [29] for SDEs driven by pure jump processes, and [17] for SDEs driven by both Gaussian and jump noises, [3] for Neutral functional SDEs and [24] for SDEs driven by a fractional Brownian motion, Li and Luo [14] and Boufoussi and Hajji [2] for stochastic delay evolution equations driven by fractional Brownian motion with Hurst parameter H∈(1/2,1) and H∈(0,1/2) in infinite dimensional space, respectively.
However, to the best of our knowledge, there is no result on the transportation inequalities for stochastic differential equations with Markovian switching. Motivated by the need of hybrid system modeling and in connection with the above discussions, it is worthwhile to develop some techniques and methods to explore the transportation inequalities for doubly perturbed stochastic differential equations with Markovian switching. To this end, in this paper, we will investigate the properties T2(C) for law of the solution of doubly perturbed stochastic equations (1.1) with Markovian switching under the L2 metric and the uniform metric. Because this kind of stochastic differential equations contain continuous-time Markov chains and doubly perturbed terms, the structure of this kind of stochastic differential equations become complex. Therefore, it is complicated to study the properties T2(C) for law of the solution of the considered equations.
The rest of this paper is organized as follows. In Section 2, we present some necessary preliminaries. In Section 3, we state and prove our main results.
Let (Ω,F,{Ft}t≥0,P) be a complete probability space with a filtration {Ft}t≥0 satisfying the usual conditions (i.e., it is increasing and right continuous, while F0 contains all P-null sets). Let {B(t)}t≥0 be a one-dimensional Brownian motion defined on the probability space (Ω,F,{Ft}t≥0,P). Let {r(t),t∈[0,+∞)} be a right continuous Markov chain on the probability space (Ω,F,{Ft}t≥0,P) taking values in a finite state space S={1,2,⋯,N}, where N is some positive integer, with generator Γ=(γi,j)N×N given by
P(r(t+△)=j|r(t)=i))={γi,j△+o(△),ifi≠j,1+γi,j△+o(△),ifi=j, |
with △>0, where γi,j≥0 is the transition rate from i to j, if i≠j; while γi,i=−∑j≠iγi,j. We assume that Markov chain r(⋅) is independent of the B(⋅). It is known that almost every sample path of r(t) is a right continuous step function with a finite number of simple jumps in any finite sub-interval of [0,+∞).
In order to obtain our main results, we impose the following three hypotheses.
Hypothesis 1. There exists some constant K>0 such that for any fixed t∈[0,+∞), i∈S and x,y∈R, the following inequality holds:
|f(t,x,i)−f(t,y,i)|2+|g(t,x,i)−g(t,y,i)|2≤K|x−y|2. |
Hypothesis 2. For each i∈S, f(⋅,0,i), g(⋅,0,i)∈L2([0,T];R) and for all t∈[0,T] and some constant ¯K>0, it follows that
|f(t,0,i)|2+|g(t,0,i)|2≤¯K. |
Hypothesis 3. |α|+|β|<1.
Lemma 2.1. If the Hypothesis 1-3 hold and the random variable x is independent of B(t), t≥0 and E|x|p<∞ for p∈N, then there exists a unique Ft-adapted solution X(t), t≥0 for (1.1) such that for any p∈N and T>0,
E(max0≤t≤T|X(t)|p)≤C, |
where C=C(T,p) and E denotes the expectation under probability measure P.
Proof. We construct the solution by iteration. Let
X0(t)=x1−α−β,0≤t<∞. |
For n≥1 define Xn+1(t) to be the unique adapted solution the following equation:
Xn+1(t)=x+∫t0f(s,Xn(s),r(s))ds+∫t0g(s,Xn(s),r(s))dB(s)+αmax0≤s≤tXn+1(s)+βmin0≤s≤tXn+1(s). | (2.1) |
Then, by view of the Theorem 3.1 of [15], we know that {Xn+1(t),t≥0} is Cauchy and X(t)=limn→∞Xn(t) is the unique Ft-adapted solution for (1.1).
Now, by using (2.1) and |α|+|β|<1, we have for any p≥2,
E(max0≤t≤T|X(t)|p)≤3p−1(11−|α|−|β|)p(E|x|p+E[max0≤t≤T|∫t0f(s,X(s),r(s))ds|p]+E[max0≤t≤T|∫t0g(s,X(s),r(s))dB(s)|p]). |
Then, using similar arguments as in the proof of the Proposition 3.5 of [33] we can obtain our desired results. The proof is complete.
Lemma 2.2. If the Hypothesis 1-3 hold and the random variable x is independent of B(t), t≥0 and E|x|2<∞, then for all 0≤s<t≤T,
E|X(t)−X(s)|2≤C1(t−s), |
where C1 is a positive constant.
Proof. For all 0≤s<t≤T, it follows from (1.1) that
X(t)−X(s)=∫tsf(u,X(u),r(u))du+∫tsg(u,X(u),r(u))dB(u)+αmax0≤u≤tX(u)+βmin0≤u≤tX(u)−αmax0≤u≤sX(u)−βmin0≤u≤sX(u). | (2.2) |
For β≥0, noticing that min0≤u≤tX(u)≤min0≤u≤sX(u), we have
X(t)−X(s)≤∫tsf(u,X(u),r(u))du+∫tsg(u,X(u),r(u))dB(u)+αmax0≤u≤tX(u)−αmax0≤u≤sX(u). | (2.3) |
Then,
|X(t)−X(s)|≤|∫tsf(u,X(u),r(u))du|+|∫tsg(u,X(u),r(u))dB(u)|+|α||max0≤u≤tX(u)−max0≤u≤sX(u)|. | (2.4) |
Next, let us consider the following two cases.
Case I. If max0≤u≤tX(u)=max0≤u≤sX(u), then we get from (2.4) that
|X(t)−X(s)|≤|∫tsf(u,X(u),r(u))du|+|∫tsg(u,X(u),r(u))dB(u)|. | (2.5) |
Case II. If max0≤u≤tX(u)>max0≤u≤sX(u), then there exists σ∈(s,t] such that X(σ)=max0≤u≤tX(u). So we get from (2.4) that
|X(t)−X(s)|≤|∫tsf(u,X(u),r(u))du|+|∫tsg(u,X(u),r(u))dB(u)|+|α||X(σ)−max0≤u≤sX(u)|≤|∫tsf(u,X(u),r(u))du|+|∫tsg(u,X(u),r(u))dB(u)|+|α||X(σ)−X(s)|≤|∫tsf(u,X(u),r(u))du|+|∫tsg(u,X(u),r(u))dB(u)|+|α|maxs≤s′<t′≤t|X(t′)−X(s′)|. | (2.6) |
Thus, we obtain that
maxs≤s′<t′≤t|X(t′)−X(s′)|2≤2(1−|α|)2(|∫tsf(u,X(u),r(u))du|2+|∫tsg(u,X(u),r(u))dB(u)|2). | (2.7) |
Then, we have
E|X(t)−X(s)|2≤2(1−|α|)2(E|∫tsf(u,X(u),r(u))du|2+E|∫tsg(u,X(u),r(u))dB(u)|2). | (2.8) |
For β<0, if min0≤u≤tX(u)=min0≤u≤sX(u), then we have also that (2.4) holds. If min0≤u≤tX(u)<min0≤u≤sX(u), then there exists δ∈(s,t] such that X(σ)=min0≤u≤tX(u). Thus, we have
|βmin0≤u≤tX(u)−βmin0≤u≤sX(u)|=|β|(min0≤u≤sX(u)−X(σ))≤|β|(X(s)−X(σ))≤|β|maxs≤s′<t′≤t|X(t′)−X(s′)|. | (2.9) |
Thus, we have
E|X(t)−X(s)|2≤2(1−|α|−|β|)2(E|∫tsf(u,X(u),r(u))du|2+E|∫tsg(u,X(u),r(u))dB(u)|2). | (2.10) |
Lastly, using the Lemma 2.1 and the Theorem 4.6.3 of [12] we can obtain our desired results.
In this section, we discuss the TCIs for the law of the solution for (1.1).
Theorem 3.1. Let the Hypothesis 1-3 hold and Px be the law of X(⋅,x), solution process for (1.1). Assume further that g is bounded by ˜g:=max0≤t≤T|g(t,x(t),r(t))|. Then the probability measure Px satisfies T2(C) on the metric space C([0,T];R) with:
(a) C=3T˜g21−|α|−|β|e6KT+24K1−|α|−|β| with the metric
d∞(γ1,γ2):=max0≤t≤T|γ1−γ2|,γ1,γ2∈C([0,T];R); |
(b) C=3T2˜g21−|α|−|β|e6KT+24K1−|α|−|β| when using the metric
d2(γ1,γ2)=(∫T0|γ1(t)−γ2(t)|2dt)1/2,γ1,γ2∈C([0,T];R). |
Proof. Let Px be the law of X(⋅,x) and Q be any probability measure such that Q≪Px. Define
˜Q:=dQdPx(X(⋅,x))P, | (3.1) |
which is a probability measure on (Ω,F). Recalling the definition of entropy and adopting a measure-transformation argument we obtain from (3.1) that
H(˜Q|P)=∫Ωln(d˜QdP)d˜Q=∫Ωln(dQdPx(X(⋅,x)))dQdPx(X(⋅,x))dP=∫C([0,T];R)ln(dQdPx)dQdPxdPx=H(Q|Px). |
Following [10], there exists a predictable process {h(t)}0≤t≤T∈R with ∫T0|h(s)|2ds<∞, P-a.s., such that
H(˜Q|P)=H(Q|Px)=12E˜Q|h(t)|2dt. |
Due to the Girsanov theorem, the process (˜B(t))t∈[0,T] defined by
˜B(t)=B(t)−∫t0h(s)ds |
is a Brownian motion with respect to {Ft}t≥0 on the probability space (Ω,F,˜Q). Consequently, under the measure ˜Q, the process {X(t,x)}t∈[0,T] satisfies
X(t)=x+∫t0f(s,X(s),r(s))ds+∫t0g(s,X(s),r(s))d˜B(s)+∫t0g(s,X(s),r(s))h(s)ds+αmax0≤s≤tX(s)+βmin0≤s≤tX(s). | (3.2) |
We now consider the solution Y (under ˜Q) of the following equation:
Y(t)=x+∫t0f(s,Y(s),˜r(s))ds+∫t0g(s,Y(s),˜r(s))d˜B(s)+αmax0≤s≤tY(s)+βmin0≤s≤tY(s). | (3.3) |
By the Lemma 2.1, under ˜Q the law of Y(⋅) is Px. Thus (X,Y) under ˜Q is a coupling of (˜Q,Px), and it follows that
[Wd22(Q,Px)]2≤E˜Q(|d2(X,Y)|2)=E˜Q(∫T0|X(t)−Y(t)|2dt),[Wd∞2(Q,Px)]2≤E˜Q(|d∞(X,Y)|2)=E˜Q(max0≤t≤T|X(t)−Y(t)|2). |
where, E˜Q denotes the expectation under probability measure ˜Q.
We now estimate the distance between X and Y with respect to d2 and d∞.
Note from (3.2) and (3.3) that
X(t)−Y(t)=∫t0[f(s,X(s),r(s))−f(s,Y(s),˜r(s))]ds+∫t0g(s,X(s),r(s))h(s)ds+∫t0[g(s,X(s),r(s))−g(s,Y(s),˜r(s))]d˜B(s)+αmax0≤s≤tX(s)−αmax0≤s≤tY(s)+βmin0≤s≤tX(s)−βmin0≤s≤tY(s). | (3.4) |
Note that
|max0≤s≤tX(s)−max0≤s≤tY(s)|≤max0≤s≤t|X(s)−Y(s)| |
and
|min0≤s≤tX(s)−min0≤s≤tY(s)|≤max0≤s≤t|X(s)−Y(s)|. |
Then, by view of (3.4) we can obtain
(1−|α|2−|β|2)E˜Q(max0≤s≤t|X(s)−Y(s)|2)≤3E˜Q(max0≤s≤t|∫s0[f(u,X(u),r(u))−f(u,Y(u),˜r(u))]du|2)+3E˜Q(max0≤s≤t|∫s0g(u,X(u),r(u))h(u)du|2)+3E˜Q(max0≤s≤t|∫t0[g(u,X(u),r(u))−g(u,Y(u),˜r(u))]d˜B(u)|2):=3(I1(t)+I2(t)+I3(t)). | (3.5) |
Firstly, applying the Hypothesis 1 and the Hölder's inequality, we have
I1(t)≤2E˜Q(max0≤s≤t|∫s0[f(u,X(u),r(u))−f(u,Y(u),r(u))]du|2)+2E˜Q(max0≤s≤t|∫s0[f(u,Y(u),r(u))−f(u,Y(u),˜r(u))]du|2)≤2KT∫t0E˜Q(max0≤u≤s|X(u)−Y(u)|2)ds+I12(t). | (3.6) |
By the Hölder's inequality, we can obtain
I12(t)≤2TE˜Q(∫t0|f(s,Y(s),r(s))−f(s,Y(s),˜r(s))|2ds)≤2T[t/ζ]∑k=0E(∫tk+1tk|f(s,Y(s),r(s))−f(s,Y(tk),r(s))|2ds+∫tk+1tk|f(s,Y(tk),r(s))−f(s,Y(tk),˜r(s))|2ds+∫tk+1tk|f(s,Y(tk),˜r(s))−f(s,Y(s),˜r(s))|2ds)=:2T[t/ζ]∑k=0[I121(t)+I122(t)+I123(t)], | (3.7) |
where t0=0≤⋯≤tk≤ttk+1≤⋯≤t[t/ζ]≤t[t/ζ]+1=t and tk+1−tk=ζ, k=0,1,⋯,[t/ζ]−1. Then, by the Hypothesis 1 and the Lemma 2.2, we have
I121(t)≤K∫tk+1tkE˜Q|Y(s)−Y(tk)|2ds≤K∫tk+1tk(s−tk)ds≤K∫tk+1tkζds, | (3.8) |
and
I123(t)≤K∫tk+1tkζds. | (3.9) |
Now, we estimate the term I122(t). Note that
I122(t)≤2E˜Q(∫tk+1tk|f(s,Y(tk),r(s))−f(s,Y(tk),r(tk))|2ds+∫tk+1tk|f(s,Y(tk),r(tk))−f(s,Y(tk),˜r(s))|2ds)=:2(I1122(t)+I2122(t)). | (3.10) |
By the Hypothesis 1-2 and the Lemma 2.2 we know that
I1122(t)=E˜Q∑i∈S∑j≠i∫tk+1tk|f(s,Y(tk),j)−f(s,Y(tk),i)|2Ir(s)=jIr(tk)=ids≤KE˜Q∑i∈S∑j≠i∫tk+1tk[1+|Y(tk)|2]Ir(tk)=iE[Ir(s)=j|Y(tk),r(tk)=i]ds≤KE˜Q∑i∈S∫tk+1tk[1+|Y(tk)|2]Iα(tk)=i[∑j≠iqij(Y(tk))(s−tk)+o(s−tk)]ds≤K∫tk+1tkO(ζ)ds, | (3.11) |
where qij(⋅) is the generator of r(t).
On the other hand, by the Hypothesis 1-2 and the following estimate (see [22,32]) that
E˜Q[I˜r(s)=j|˜r(tk)=i1,r(tk)=i,X(tk),Y(tk)]=∑l∈SE˜Q[I˜r(s)=j,Ir(s)=l|˜r(tk)=i1,r(tk)=i,X(tk),Y(tk)]=∑l∈Sq(i1,i)q(j,l)(s−tk)+o(s−tk)=O(ζ), |
we have
I2122(t)=E˜Q∑i∈S∑j≠i∫tk+1tk|f(s,Y(tk),j)−f(s,Y(tk),i)|2I˜r(s)=jIr(tk)=ids≤KE˜Q∑i∈S∑j≠i∫tk+1tk[1+|Y(tk)|2]I˜r(tk)=i1Ir(tk)=i×E˜Q[I˜r(s)=j|˜r(tk)=i1,r(tk)=i,X(tk),Y(tk)]ds≤K∫tk+1tkO(ζ)ds. | (3.12) |
From (3.7)–(3.12), letting ζ→0, we have I12(t)→0. Thus, combined with (3.6) we obtain
|I1(t)|≤2KT∫t0E˜Q(max0≤u≤s|X(u)−Y(u)|2)ds. | (3.13) |
Since h∈L2([0,T];R), by the boundedness of g and the Hölder's inequality, we can obtain
E˜Q(max0≤s≤t|∫s0g(u,X(u),r(u))h(u)du|2)≤T˜g2∫t0E˜Q|h(s)|2ds. | (3.14) |
Next, we estimate the term I3(t). By using the Burkholder-Davis-Gundy's inequality and the assumptions on g we know that
I3(t)≤4E˜Q∫t0|g(s,X(s),r(s))−g(s,Y(s),˜r(s))|2ds≤8E˜Q(∫t0|g(s,X(s),r(s))−g(s,Y(s),r(s))|2ds)+8E˜Q(∫t0|g(s,Y(s),r(s))−g(s,Y(s),˜r(s))|2ds)≤8K∫t0E˜Q(max0≤u≤s|X(u)−Y(u)|2)ds+I32(t). |
Then, being similar to I1(t), we also have
I3(t)≤8K∫t0E˜Q(max0≤u≤s|X(u)−Y(u)|2)ds. | (3.15) |
Thus, combining (3.5), (3.13), (3.14) with (3.15) we have
E˜Q(max0≤s≤t|X(s)−Y(s)|2)≤6KT+24K1−|α|−|β|∫t0E˜Q(max0≤u≤s|X(u)−Y(u)|2du+3T˜g21−|α|−|β|∫t0E˜Q|h(s)|2ds. | (3.16) |
Now, the Gronwall's lemma implies that for any t>0,
E˜Q(max0≤s≤t|X(s)−Y(s)|2)≤3T˜g21−|α|−|β|e6KT+24K1−|α|−|β|∫t0E˜Q|h(s)|2ds, |
which implies that
E˜Q|X(t)−Y(t)|2≤3T˜g21−|α|−|β|e6KT+24K1−|α|−|β|∫t0E˜Q|h(s)|2ds. |
Hence, we may write that
d2∞(X,Y)≤3T˜g21−|α|−|β|e6KT+24K1−|α|−|β|∫T0E˜Q|h(s)|2ds |
and
[Wd∞2(Q,Px)]2≤2CT,KH(Q|Px) |
with CT,K=3T˜g21−|α|−|β|e6KT+24K1−|α|−|β|.
Analogously for the metric d2, we have by the Fubini's theorem
[Wd22(Q,Px)]2≤E˜Q(∫T0|X(t)−Y(t)|2dt)=∫T0E˜Q(|X(t)−Y(t)|2)dt≤3T˜g21−|α|−|β|e6KT+24K1−|α|−|β|∫T0∫t0E˜Q|h(s)|2dsdt=3T˜g21−|α|−|β|e6KT+24K1−|α|−|β|∫T0E˜Q|h(s)|2(∫Ts1⋅dt)ds≤3T2˜g21−|α|−|β|e6KT+24K1−|α|−|β|∫T0E˜Q|h(s)|2ds. |
Thus, we can obtain
[Wd22(Q,Px)]2≤2CT,KH(Q|Px) |
with CT,K=3T2˜g21−|α|−|β|e6KT+24K1−|α|−|β|. The proof is complete.
Remark 3.1. There are many interesting applications of the TCIs, e.g., in Tsirel'son-type inequality and Hoeffding-type inequality, see [9,30,31], and in concentration of empirical measure [17,29].
We would like to thank the referees for the careful reading of the manuscript and for their valuable suggestions. This research is supported by the National Natural Science Foundation of China (No.11901058 and No.62076164).
The authors declare that they have no conflict of interest.
[1] |
S. Bobkov, F. Götze, Exponential integrability and transportation cost related to logarithmic Sobolev inequalities, J. Funct. Anal., 163 (1999), 1–28. doi: 10.1006/jfan.1998.3326
![]() |
[2] |
B. Boufoussi, S. Hajji, Transportation inequalities for neutral stochastic differential equations driven by fractional Brownian motion with Hurst parameter lesser than 1/2, Mediterr. J. Math., 14 (2017), 192. doi: 10.1007/s00009-017-0992-9
![]() |
[3] |
J. Bao, F. Wang, C. Yuan, Transportation cost inequalities for neutral functional stochastic equations, Z. Anal. Anwend., 32 (2013), 457–475. doi: 10.4171/ZAA/1494
![]() |
[4] |
P. Carmona, F. Petit, M. Yor, Some extensions of the arc sine law as partial consequences of the scaling property of Brownian motion, Probab. Theory Relat. Fields, 100 (1994), 1–29. doi: 10.1007/BF01204951
![]() |
[5] | P. Carmona, F. Petit, M. Yor, Beta variables as times spent in [0,∞) by certain perturbed Brownian motions, J. Lond. Math. Soc., 58 (2016), 239–256. |
[6] |
L. Chaumont, R. A. Doney, Some calculations for doubly perturbed Brownian motion, Stochastic Process. Appl., 85 (2000), 61–74. doi: 10.1016/S0304-4149(99)00065-4
![]() |
[7] | B. Davis, Brownian motion and random walk perturbed at extrema. Probab. Theory Relat. Fields., 113 (1999), 501–518. |
[8] |
R. A. Doney, T. Zhang, Perturbed Skorohod equations and perturbed reflected diffusion processes, Ann. Inst. H. Poincaré Probab. Statist., 41 (2005), 107–121. doi: 10.1016/j.anihpb.2004.03.005
![]() |
[9] | H. Djellout, A. Guilin, L. Wu, Transportation cost-information inequalities for random dynamical systems and diffsions, Ann. Probab., 32 (2004), 2702–2732. |
[10] | G. Da Prato, J. Zabczyk, Stochastic Equations in Infinite Dimonsionals, Cambridge: Cambridge University Press, 1992. |
[11] |
D. Feyel, A. S. Üstünel, The Monge-Kantorovitch problem and Monge-Ampère equation on Wiener space, Probab. Theory Relat. Fields., 128 (2004), 347–385. doi: 10.1007/s00440-003-0307-x
![]() |
[12] | A. Friedman, Stochastic Differential Equations and Applications, Vols. I and II, New York: Academic Press, 1975. |
[13] | L. Hu, Y. Ren, Doubly perturbed neutral stochastic functional equations, J. Comput. Appl. Math., 31 (2009), 319–326. |
[14] |
Z. Li, J. W. Luo, Transportation inequalities for stochastic delay evolution equations driven by fractional Brownian motion, Front. Math. China, 10 (2015), 303–321. doi: 10.1007/s11464-015-0387-9
![]() |
[15] |
D. Liu, Y. Yang, Doubly perturbed neutral diffusion processes with Markovian switching and Poisson jumps, Appl. Math. Lett., 23 (2010), 1141–1146. doi: 10.1016/j.aml.2009.07.004
![]() |
[16] |
J. W. Luo, Doubly perturbed jump-diffusion processes, J. Appl. Math. Anal., 351 (2009), 147–151. doi: 10.1016/j.jmaa.2008.09.024
![]() |
[17] |
Y. Ma, Transportation inequalities for stochastic differential equations with jumps, Stochastic Process. Appl., 120 (2010), 2–21. doi: 10.1016/j.spa.2009.09.012
![]() |
[18] |
W. Mao, L. J. Hu, X. R. Mao, Approximate solutions for a class of doubly perturbed stochastic differential equations, Adv. Differ. Equations, 2018 (2018), 37. doi: 10.1186/s13662-018-1490-5
![]() |
[19] |
J. Norris, L. Rogers, D. Williams, Self-avoiding random walk: A Brownian motion model with local time drift, Probab. Theory Relat. Fields, 74 (1987), 271–287. doi: 10.1007/BF00569993
![]() |
[20] |
F. Otto, C. Villani, Generalization of an inequality by Talagrand and links with the logarithmic Sobolev inequality, J. Funct. Anal., 173 (2000), 361–400. doi: 10.1006/jfan.1999.3557
![]() |
[21] |
S. Pal, Concentration for multidimensional diffusions and their boundary local times, Probab. Theory Relat. Fields, 154 (2012), 225–254. doi: 10.1007/s00440-011-0368-1
![]() |
[22] |
B. Pei, Y. Xu, G. Yin, Averaging principles for SPDEs driven by fBm with random delay modulated by two-time-scale Markovian switching processes, Stoch. Dynam., 18 (2018), 1850023. doi: 10.1142/S0219493718500235
![]() |
[23] | M. Perman, W. Werner, Perturbed Brownian motions, Probab. Theory Relat. Fields, 108 (1997), 357–383. |
[24] | B. Saussereau, Transportation inequalities for stochastic differential equations driven by a fractional Brownian motion, Bernoulli, 18 (2012), 1–23. |
[25] |
M. Talagrand, Transportation cost for Gaussian and other product measures, Geom. Funct. Anal., 6 (1996), 587–600. doi: 10.1007/BF02249265
![]() |
[26] |
A. S. Üstünel, Transport cost inequalities for diffusions under uniform distance, Stochastic Anal. Relat. Top., 22 (2012), 203–214. doi: 10.1007/978-3-642-29982-7_9
![]() |
[27] | F. Y. Wang, Transportation cost inequalities on path spaces over Riemannian manifolds, Illinois J. Math., 46 (2002), 1197–1206. |
[28] |
F. Y. Wang, Probability distance inequalities on Riemannian manifolds and path spaces, J. Funct. Anal., 206 (2004), 167–190. doi: 10.1016/S0022-1236(02)00100-3
![]() |
[29] | L. Wu, Transportation inequalities for stochastic differential equations of pure jumps, Ann. Inst. Henri Poincaré Probab. Stat., 46 (2010), 465–479. |
[30] | L. Wu, Z. Zhang, Talagrand's T2-transportation inequality w.r.t. a uniform metric for diffusions, Acta Math. Appl. Sin. Engl. Ser., 20 (2004), 357–364. |
[31] |
L. Wu, Z. Zhang, Talagrand's T2-transportation inequality and log-Sobolev inequality for dissipative SPDEs and applications to reaction-diffusion equations, Chinese Ann. Math. Ser. B, 27 (2006), 243–262. doi: 10.1007/s11401-005-0176-y
![]() |
[32] |
G. Yin, C. Zhu, Properties of solutions of stochastic differential equations with continuous-state-dependent switching, J. Differ. Equations, 249 (2010), 2409–2439. doi: 10.1016/j.jde.2010.08.008
![]() |
[33] |
C. Zhu, G. Yin, On strong feller, recurrence, and weak stabilization of regime-switching diffusions, SIAM J. Control Optim., 48 (2009), 2003–2031. doi: 10.1137/080712532
![]() |
1. | Xiaocui Ma, Haitao Yue, Fubao Xi, The averaging method for doubly perturbed distribution dependent SDEs, 2022, 189, 01677152, 109588, 10.1016/j.spl.2022.109588 |