1 Introduction

Let(Ω,F,P)be a basic probability space.T ∈ (0,+∞)is a finite time to be con firmed.Let BH=(BH1(t),···,BHm(t)),t ∈ [0,T]be an m-dimensional fractional Brownian motion(fBm,for short)of Hurst parameter H∈(1/2,1).It is allowed that the Hurst parameter H may be different fromtoexcept notation for each i/=j.In other words,,j=1,···,m,are independent,continuous,and mean zero Gaussian processes with the following covariance

where δ is the Kronecker symbol.For each j=1,···,m,and t ∈ [0,T],letbe the σalgebra by,0≤s≤t,augmented by all the P-zero measurable events.We denote the corresponding filtration byLet W=(W1(t),···,Wm(t)),t∈ [0,T]be an F-adapted m-dimensional standard Brownian motion(sBm,for short).We say that W=(W1(t),···,Wm(t))is the underlying standard Brownian motion corresponding to BH=(BH1(t),···,BHm(t)),t∈ [0,T]if

where

and

We consider the following backward stochastic differential equation

where ξ is given terminal value,f and g are the given(random)generators.To solve this equation is to find a pair of adapted processes(Yt,Zt)0≤t≤T,satisfying the above equation(1.1).

where⋄denotes the Wick product.Furthermore,we have

Because the fractional Brownian motions are not semimartingales except the case H=1/2,it can be used to describe and explain some natural or society phenomenons,such as to model hydrology,climatology,signal processing,network traffic analysis,control theory, finance as well as various other fields.This makes the stochastic analysis for fractional Brownian motions challenging and fascinating.In recent years,the stochastic integral with respect to the fBm had been investigated from different perspectives by many authors[1,9,10,20,26],and so on.For the Malliavin calculus with respect to fBm and applications,thanks to the contribution of[8,14,21,22]and references there.Hu and Peng in[17]had studied the backward stochatic differential equations driven by fractional Brownian motions.They studied the general and linear BSDEs driven by fractional Brownian motions.Shortly afterwards Fei,Xia,and Zhang in[11]had analogous results as Hu and Peng in[17]to the BSDEs driven by both standard and fractional Brownian motions.In both articles,the pair of solutions of BSDEs(Yt,Zt;0≤t≤T)relied on the simple process ηt= η0+b(t)+where b and σ are determinate function in L2([0,T]).In this article,we will cancel this dependency.

广东省9家“三甲”医院药师对互联网药事服务认知情况的调查分析 ……………………………………… 汪小惠等（2）：159

Especially,in stochastic optimal control theory for controlled system,BSDEs as the adjoint equations corresponding to the state equations to describe optimal control problem are essential,for example,[2–6,13,18,19,24].Recently,Han,Hu,and Song[12]obtained a stochastic maximum principle for a stochastic control problem and their adjoint backward stochastic differential equation is linear driven by the fractional Brownian motions and its underlying standard Brownian motions.They gave the explicit form solution.Furthermore,we will show the local existence and uniqueness of solutions to the general nonlinear BSDEs driven by the fractional Brownian motions and its underlying standard Brownian motions.Diehl and Friz in[7]had studied some similar equations,which used a rough paths point view.The results of Diehl and Friz were less general than those proved in this article as g were not allowed to depend to Z.

就在他们要用担架把我抬向某个地方的时候，我奋力地从那担架上滚落下来。此时此刻，我的第一愿望是快快快！我要尽量地快些到那个黑乎乎的矿井底下。我放心不下。我要找到小六子啊。

This article is arranged as follows.In Section 2,we introduce some basic notations concerning the fractional Brownian motions,standard Brownian motions,and Mallivian calculus taken from[8,10,14,21].In Section 3,we will give some results concerning stochastic differential equations driven by both fractional and standard Brownian motions.We prove the generalization of the Itˆo formula of an Itˆo type process involving the stochastic integral with respect to both standard and fractional Brownian motions.Moreover,we obtain the product formula as the corollary.In Section 4,we prove the local existence and uniqueness of the solutions to the above BSDEs(1.1)under some reasonable assumptions.

2 Preliminaries on fBm,sBm and Malliavin Calculus

We denote bythe space of Rn-valued,F-adapted random variable F such thatthe space of F-adapted random variable F such that

Let{ρ1,···,ρk,···}be an orthogonal basic of L2([0,T])such that ρk,k=1,2,···,are smooth functions on[0,T].Let PTrepresent the set of all polynomials of standard Brownian motion over interval[0,T].Namely,PTcontains all elements of the form

where f is a polynomial of n variables.The Malliavin derivative Dtof a polynomial functional F of standard Brownian motion is defined as

For any F∈PT,we denote the following norm

选择2016年1月——2017年12月于我院分娩的100例新生儿作为研究对象,本组新生儿均于我院顺利分娩,均为单胎,其中男56例,女44例,年龄1~30d,平均(16.35±2.31)d,出生体重2500~4200g,平均(3000.23±45.52)g;顺产51例,剖宫产49例。

Let ξ and η be two continuous functions on[0,T].Define a Hilbert scalar product

Let D1,2denote the Banach space obtained by completing PTunder the norm ‖ ·‖1,2.

后来胡马强多次邀请范峥峥去他那里当办公室主任，并在一次为范峥峥设的酒宴上下了药，范峥峥醉酒后被搀扶走进胡宅的那天晚上，正是贾鹏飞等人把我栽到胡马强院子之时。

where φ :[0,T]→ R+satisfy

When ξ= η,we define the Hilbert norm by

基本公共服务均等化是指政府在不同时期按照同一标准为社会公众提供基本的、大致均等的基本公共服务。本文采用基本公共服务均等化指数来测度我国基本公共服务均等化供给水平，该指数的创建主要使公共服务均等化程度能够在一定程度得到量化，由于它来源于地区间财政支出，因此与前文的机理模型中的影响因素息息相关。由式（16）给出。测度所用原始数据来源于《中国统计年鉴》和《中国城市统计年鉴》。

We denote Θtas the completion of the continuous functions under this Hilbert norm.

Let{η1,···,ηk,···}be an orthogonal basic of ΘTandbe the set of all polynomials of fractional Brownian motion over interval[0,T].Namely,contains all elements of the form

where g is a polynomial of n variables.For any G∈,the Malliavin derivativeof a polynomial functional G of fractional Brownian motion of Hurst parameter H is defined by

Introduce another Malliavin derivative

For any G∈,we denote the following norm

Let DH,1,2denote the Banach space obtained by completingunder the norm ‖ ·‖H,1,2.We can certainly define the higher derivatives and the space Dk,pand space DH,k,p.But we only need the space D1,2and DH,1,2in this article.

Recall the operator (see[14],(5.32))

whereand Γ(α)is the gamma function.If F is a nice function on[0,T](see[14]),we have

where BH(t)is a fractional Brownian motion and W(t)is the corresponding to BH(t),underlying standard Brownian motion.

Next,let us recall that the framework of stochastic integral with respect to fractional Brownian motion.The following result is well known in[10,14,15].

Proposition 2.1 If Fsis a continuous stochastic process such that

then,the Itˆo-type stochastic integral can be defined under πn:0=t0 ＜ t1 ＜ ···＜ tn=T of the partition of the internal[0.T]:

Backward stochastic differential equations have a variety of applications in fields such as stochastic optimal control theory,mathematical finance,probability interpretation of solutions of quasi-linear partial differential equations and so on.The general nonlinear backward stochastic differential equations with respect to standard Brownian motions were first introduced by Pardoux and Peng[25].Peng recently gave a survey[23]on the developments in the theory of nonlinear BSDEs during the past 20 years,including the existence and uniqueness of the solutions,nonlinear Feynman-Kac formula,nonlinear expectation and many other results in BSDEs theory and their applications to dynamic pricing and hedging in an incomplete financial market.

and

Let La(Ω;R)be the family of stochastic processes on[0,T]such that La(Ω;R)if F is R-valued,F-adapted random variable,satisfying condition(2.4).

We need the following integration by parts formula.

Proposition 2.2(see[8],Theorem 3.15)Let F∈D1,2and let(u(t),t∈[0,T])be a Itˆo integrable stochastic process such that(Fu(t),t∈[0,T])is It integrable.Then,

Another proposition is well known from(see[14],Theorem 10.2).

Proposition 2.3 Let F∈DH,1,2and let(f(t),t∈[0,T])be(not necessarily adapted)stochastic process such that f satisfies condition 2.4.Then,

3 Itˆo Formula With Respect to Both fBm and sBm

In this section,we will give some important results concerning stochastic differential equations driven by both fractional and standard Brownian motions that will be used in the sequel.

Now,we give the generalization of the Itˆo formula of an Itˆo type process involving the stochastic integral with respect to both fractional and standard Brownian motions.

Theorem 3.1 Let Gt∈ (Ω;R),Ft∈ La(Ω;R)and Nt∈ (Ω;R),t∈ [0,T],be real valued stochastic processes.Suppose thatAssume also thatare continuously differential with respect to(s,t)∈ [0,T]× [0,T]for almost all ω ∈ Ω.Denote

where X0is a constant.Let f be a function having the first continuous derivative with respect to t and twice continuous derivative with respect to x and suppose that these derivatives are bounded.Then,we have

普通患儿一天蓝光照射治疗仅需几个小时，却要占用24小时床位，“其实没必要。”于是，中心将其治疗统一实行门诊日间照射治疗模式，孩子既可以回家，妈妈也可以喂奶，还能节省医院的水、电、人员支出等住院消耗。

for 0≤t≤T.

Proof Let πnbe a partition defined as follows:

Then,we have

羊肉肉质细腻，是人们喜欢的食物，因此市场的销售量逐渐增加。近年，随着市场需求增大，养羊业也在不断发展，肉羊养殖经济效益大。但也导致传染性强的疾病频发，疾病一旦发生，传染率极高，给养羊户造成巨大的经济损失，尤其是在养殖中容易出现诸多问题。因此，需要结合具体的实际情况和养殖业的实际特点，对肉羊养殖管理措施进行系统分析，使后续的养殖技术得以完善。随着肉羊养殖环境的快速变化，传统模式的肉羊疾病防治模式已经难以充分适应现代条件下的肉羊疾病防治要求，因此，很多肉羊饲养团队都将疾病防治工作的研究设计作为工作的重点[1]。

By the mean value theorem,the first sum converges to

in L2.Now,we calculate the second sum.Using Taylor′s formula and expanding each term in the sum to the second order,we obtain

目前，国内关于志愿者服务的研究主要的关注点在于志愿服务在国内的发展历程、发展阻碍因素与解决办法以及如何建立健全志愿服务机制、完善关于志愿服务的法律法规和如何激励参与者积极参与志愿服务等等问题。而笔者的论述主要以大学生为调查对象，从大学生个人发展意义的角度出发来论述参与志愿服务活动的必要性。所谓志愿服务的含义：就是参与志愿服务活动的个人自愿贡献自己的时间及精力，在不计报酬的情况下，所从事或参加的服务他人、奉献社会的社会公益实践活动。

whereNow,the proof will be decomposed in several steps.

Step 1 Let us show that

按照中华医学会耳鼻咽喉科学分会杭州标准[9]，治愈8例（40%），有效9例（45%），无效3例（15%），总有效率为85%。H-UPPP手术前后AHI、最低血氧饱和度（The lowest oxyhemoglobin saturation,LSaO2）、ESS和鼾声量表评分均明显改善，其中除AHI指数手术前后变化P值小于0.05外，其它指标变化P值均小于0.01。20例患者手术前后的BMI指数发生变化，而且具有统计学意义，具体数据变化参见表1。

in L2as n→+∞.

对于企业来说，安全生产事关经济效益和生产效率的提高，事关企业的可持续、健康发展。对于员工来说，安全事关生命，也是企业对员工的基本承诺。而对于海外项目而言，安全已经成为队伍能否进入国际服务市场首先必须跨越的一道门槛；没有高标准的安全管理体系和良好的业绩就很难通过招投标资格预审，更别说拿到项目合同。

In this section,we will consider the local existence and uniqueness of solutions to the following BSDEs driven by the fractional Brownian motions and the underlying standard Brownian motions.

in L2as n→+∞.We can compute the following term by formulas(2.6)and(2.7).Drut=0,if r＞ t,∀ ut∈(Ω;R)(see[8],Corollary 3.13).Thus,we have

The incrementcan be decomposed into

这些汇集了国内外艺术家智慧和才能的雕塑艺术，形态各异，风格多样，各领风骚，日月千秋。它们以当代的艺术视角、切中民勤的历史文化、地域特性和当代精神风貌等多层面、深刻生动地阐发出了民勤特有的文化魅力和勃勃生机。

where tk+1− tk → 0,and C is a constant independent of the partition π that may differ from line to line.

Clearly,the contribution of I14,I15,and I16to limit(3.3)is zero.We have

By the assumptions,we know that the second derivative of f with respect to its second variable is bounded.Thus,we have

Therefore,by the isometry formula,it is easy to get

in L2as n→+∞.This completes the proof of the theorem.

Next,we provide the product formula to stochastic differential equations with respect to both fractional Brownian motions and standard Brownian motions.It is not hard to obtain the following corollary that the function f(t,x,y)=xy under the above theorem.

Corollary 3.2 Let Xtand Ytbe two processes satisfying

and

音乐剧(戏剧)是演员在舞台上的表演，观众在舞台下只能从某一个角度观看，并且距离越远，越看不清楚演员的动作。因此，作为音乐剧(戏剧)演员舞台表演，要求动作的幅度一定要大，需要做得特别夸张，才能让全场的观众看的清楚、明白。

where stochastic processes Fi(t),Gi(t),Ni(t),i=1,2,satisfy the assumptions of Theorem 3.1,and X0and Y0are constants.Then,

which may be written formally as

4 Local Existence and Uniqueness of Solutions to BSDEs

Step 2 We will show that

从商品蛋鸡效益来看，2018年前8个月鸡蛋平均出厂价为7.69元/千克，同比上升35.10%。平均鸡蛋成本 7.35元/千克，同比增加2.90%。平均淘汰鸡价格为16.21元/只，同比上升15.80%。只鸡赢利13.63元。

where T ∈ (0,+∞)is bounded time to be con firmed.The generator f:[0,T]×Rn×Rn×m →Rn and g:[0,T]×Rn×Rn×m → Rn×mare given(random)functions.ξ∈ (Ω;Rn)is a given terminal variable.Without lose of generality,we let n=m=1.

Let(Ω;R)be the completion of (Ω;R)under the following β-norm:

We have the following assumptions on the generator f and g.

Assumption(H):

(H1)f(t,y,z)∈ (Ω;R)is continuously differentiable,and mean square bounded.Its Malliavin derivatives exists,satisfyingMoreover,the Lipschitz conditions hold for f and That is,

(H2)g(t,y,z)∈ La(Ω;R)is twice continuously differentiable with respect to y and z,and itsfirst and second-order Malliavin derivatives exist,all of which are mean square bounded.Moreover,there exists a constant C＞0 such that the following inequality hold:

Definition 4.1 A solution to BSDE(4.1)is a pair of stochastic processes(Yt,Zt)0≤t≤T ∈and the pair of processes satisfies

Theorem 4.2 Suppose that(H1)and(H2)hold.Then,there exists uniquely a pair of solutionto BSDE(4.1)with any terminal value

To prove Theorem 4.2,we consider firstly the simple case of f(t,yt,zt)=f(t)and g(t,yt,zt)=g(t).That is,BSDE(4.1)has the following form

Lemma 4.3 Assume that ft∈ and gt∈ La(Ω;R).Then,there exists uniquely a pair of solutionof BSDE(4.2).

Proof The proof of the uniqueness is obvious.We omit it.For the existence of the solution,we show the following statement.Let

It is easy to see that Mtis a Ft−martingale.With the martingale representative theorem,there exists a unique stochastic process zt∈ for all 0≤ t≤ T.Moreover,

Next,we define ytas follows:

It is easy to see that y(t)∈Furthermore,is the solution of BSDE(4.2).

For each(yt,zt)→ (Yt,Zt),we know that BSDE(4.1)has a pair of solutionby Lemma 4.3.Let Γ be the map defined as follows,

To complete the proof of Theorem 4.2,what we need is to prove that the map Γ is contractive.Suppose thatandare solutions of BSDE(4.1),corresponding to(yt,zt)=respectively.Then,we have

Let

We have

Now,the proof will be decomposed in several steps.

Step 1 The integral form can be written by

It is well-known that DrXt=0,if r＞ t,∀Xt∈ We have the following equations by(2.6),(2.7),and(2.3).

The Malliavin derivative can be expressed with(2.2)and(2.3)by

Thus,

Step 2 By formula(3.4),we have

Again using the product formula,we have

Taking expectation,we obtain

So,we have

Step 3 Firstly,we calculate J1+J2+J4.

Next,letand we have the estimate about J3as follows.C is a constant,independent on T and β,which may be different from line to line.

Now,we can group

Integrate from 0 to T for we have

We also have

Summarizing with Assumption(H),we obtain

where

So,we only need to choose T small enough and β big enough,satisfying that CQ ＜ 1.Then,we complete the proof of Theorem 4.2 by using of contractive mapping principle.

References

[1]Alòs E,Nualart D.Stochastic integration with respect to the fractional brownian motion.Stoch Stoch Rep,2003,75(3):129–152

[2]Bender C.Explicit solutions of a class of linear fractional BSDEs.Systems Control Lett,2005,54(7):671–680

[3]Bensoussan A.Lectures on Stochastic Control.Lecture Notes in Math.Berlin-New York:Springer,1982

[4]Biagini F,Hu Y,Øksendal B,Sulem A.A stochastic maximum principle for processes driven by fractional brownian motion.Stochastic Process Appl,2002,100:233–253

[5]Bismut J M.Conjugate convex functions in optimal stochastic control.J Math Anal Appl,1973,44:384–404

[6]Bismut J M.An introductory approach to duality in optimal stochastic control.SIAM Rev,1978,20(1):62–78

[7]Diehl J,Friz P.Backward stochastic differential equations with rough drivers.Ann Probab,2012,40(4):1715–1758

[8]Di Nunno G, Øksendal B,Proske F.Malliavin calculus for Lévy processes with applications to finance.Berlin:Springer-Verlag,2009

[9]Dai W,Heyde C C.Itˆo formula with respect to fractional brownian motion and its application.J Appl Math Stochastic Anal,1996,9(4):439–448

[10]Duncan T,Hu Y,Pasik-Duncan B.Stochasticcalculus for fractional brownian motion I.theory.SIAM J Control Optim,2000,38(2):582–612

[11]Fei W,Xia D,Zhang S.Solutions to BSDES driven by both standard and fractional brownian motions.Acta Math Appl Sin Engl Ser,2013,29(2):329–354

[12]Han Y,Hu Y,Song J.Maximum principle for general controlled systems driven by fractional brownian motions.Appl Math Optim,2013,67(2):279–322

[13]Haussmann U G.General necessary conditions for optimal control of stochastic system.Math Programm Stud,1976,6:34–48

[14]Hu Y.Integral transformations and anticipative calculus for fractional brownian motions.Mem Amer Math Soc,2005,175(825)

[15]Hu Y,Øksendal B.Fractional white noise calculus and applications to finance.In fin Dimens Anal Quantum Probab Relat Top,2003,6(1):1–32

[16]Hu Y,Øksendal B.Partial information linear quadratic control for jump diffusions.SIAM J Control Optim,2008,47(4):1744–1761

[17]Hu Y.Peng S.Backward stochastic differential equation driven by fractional brownian motion.SIAM J Control Optim,2009,48(3):1675–1700

[18]Hu Y,Zhou X.Stochastic control for linear systems driven by fractional noises.SIAM J Control Optim,2005,43(6):2245–2277

[19]Kushner H J.Necessary conditions for continuous parameter stochastic optimization problems.SIAM J Control,1972,10:550–565

[20]Lin S.Stochastic analysis of fractional Brownian motions.Stochastics Stochastics Rep,1995,55(1):121–140

[21]Nualart D.The Malliavin calculus and related topics.Berlin:Springer,2006

[22]Nualart D,Rascanu S.Differential equations driven by fractional brownian motion.Collect Math,2002,53(1):55–81

[23]Peng S.Backward stochastic differential equation nonlinear expectation and their applications//Proceedings of the International Congress of Mathematicians Hindustan Book Agency.India:New Delhi,2010:393–432

[24]Peng S.A general stochastic maximum principle for optimal control problems.SIAM J Control Optim,1990,28(4):966–979

[25]Pardoux E,Peng S.Adapted solution of a backward stochastic differential equation.Systems Control Lett,1990,14(1):55–61

[26]Wu L,Ding Y.Wavelet-based estimator for the Hurst parameters of fractional Brownian sheet.Acta Mathematica Scientia,2017,37B(1):205–222