Loading [MathJax]/jax/output/SVG/jax.js
Research article

Approximation of solutions of multi-dimensional linear stochastic differential equations defined by weakly dependent random variables

  • Received: 30 October 2016 Accepted: 23 May 2017 Published: 22 June 2017
  • It is well-known that under suitable conditions there exists a unique solution of a ddimensional linear stochastic differential equation. The explicit expression of the solution, however, is not given in general. Hence, numerical methods to obtain approximate solutions are useful for such stochastic di erential equations. In this paper, we consider stochastic difference equations corresponding to linear stochastic differential equations. The difference equations are constructed by weakly dependent random variables, and this formulation is raised by the view points of time series. We show a convergence theorem on the stochastic difference equations.

    Citation: Hiroshi Takahashi, Ken-ichi Yoshihara. Approximation of solutions of multi-dimensional linear stochastic differential equations defined by weakly dependent random variables[J]. AIMS Mathematics, 2017, 2(3): 377-384. doi: 10.3934/Math.2017.3.377

    Related Papers:

    [1] Xiao Qi, Mejdi Azaiez, Can Huang, Chuanju Xu . An efficient numerical approach for stochastic evolution PDEs driven by random diffusion coefficients and multiplicative noise. AIMS Mathematics, 2022, 7(12): 20684-20710. doi: 10.3934/math.20221134
    [2] Xiaoming Wang, Muhammad W. Yasin, Nauman Ahmed, Muhammad Rafiq, Muhammad Abbas . Numerical approximations of stochastic Gray-Scott model with two novel schemes. AIMS Mathematics, 2023, 8(3): 5124-5147. doi: 10.3934/math.2023257
    [3] Fawaz K. Alalhareth, Seham M. Al-Mekhlafi, Ahmed Boudaoui, Noura Laksaci, Mohammed H. Alharbi . Numerical treatment for a novel crossover mathematical model of the COVID-19 epidemic. AIMS Mathematics, 2024, 9(3): 5376-5393. doi: 10.3934/math.2024259
    [4] Yuyan Wei, Xili Tan, Peiyu Sun, Shuang Guo . Weak and strong law of large numbers for weakly negatively dependent random variables under sublinear expectations. AIMS Mathematics, 2025, 10(3): 7540-7558. doi: 10.3934/math.2025347
    [5] Masataka Hashimoto, Hiroshi Takahashi . On the rate of convergence of Euler–Maruyama approximate solutions of stochastic differential equations with multiple delays and their confidence interval estimations. AIMS Mathematics, 2023, 8(6): 13747-13763. doi: 10.3934/math.2023698
    [6] Snezhana Hristova, Kremena Stefanova . p-moment exponential stability of second order differential equations with exponentially distributed moments of impulses. AIMS Mathematics, 2021, 6(3): 2886-2899. doi: 10.3934/math.2021174
    [7] Dennis Llemit, Jose Maria Escaner IV . Value functions in a regime switching jump diffusion with delay market model. AIMS Mathematics, 2021, 6(10): 11595-11609. doi: 10.3934/math.2021673
    [8] Clara Burgos, Juan Carlos Cortés, Elena López-Navarro, Rafael Jacinto Villanueva . Probabilistic analysis of linear-quadratic logistic-type models with hybrid uncertainties via probability density functions. AIMS Mathematics, 2021, 6(5): 4938-4957. doi: 10.3934/math.2021290
    [9] Yongxiang Zhu, Min Zhu . Well-posedness and order preservation for neutral type stochastic differential equations of infinite delay with jumps. AIMS Mathematics, 2024, 9(5): 11537-11559. doi: 10.3934/math.2024566
    [10] Chunli You, Linxin Shu, Xiao-bao Shu . Approximate controllability of second-order neutral stochastic differential evolution systems with random impulsive effect and state-dependent delay. AIMS Mathematics, 2024, 9(10): 28906-28930. doi: 10.3934/math.20241403
  • It is well-known that under suitable conditions there exists a unique solution of a ddimensional linear stochastic differential equation. The explicit expression of the solution, however, is not given in general. Hence, numerical methods to obtain approximate solutions are useful for such stochastic di erential equations. In this paper, we consider stochastic difference equations corresponding to linear stochastic differential equations. The difference equations are constructed by weakly dependent random variables, and this formulation is raised by the view points of time series. We show a convergence theorem on the stochastic difference equations.


    1. Introduction

    Stochastic differential equations have been used to describe stochastic models in many areas. Under suitable conditions, a stochastic differential equation has a unique solution, and the representation of the solution is given by a stochastic integral. To express the explicit solutions of stochastic differential equations, numerical methods on stochastic differential equations have been well established. For example, we refer [1] and [3].

    In the theory of mathematical finance, Black-Scholes type stochastic differential equation is an important example. Let {W(t)} be a standard one-dimensional Wiener process. For fixed constants μR and σ>0, Black-Scholes model is given by the following stochastic differential equation:

    dX(t)=X(t)(νdt+σdW(t)), 0tT, (1.1)

    where T is a maturity. Fixing nN, we observe the stochastic differential equation at times {tk=kT/n,k=1,2,,n}. To simplify the problem, we consider the case where ν=0. By the Euler-Maruyama scheme, we can consider an approximation of the solution of (1.1) such that

    X(tk)X(tk1)σX(tk1){W(tk)W(tk1)}, k=1,2,,n,

    where means nearly equal to. We rewrite the approximation above such that

    σ{W(tk)W(tk1)}X(tk)X(tk1)X(tk1).

    In mathematical finance, we regard the difference W(tk)W(tk1) as a rate of returns. For the Black-Scholes model, the difference is given by a sequence of independent and identically distributed random variables. If we consider a stochastic difference equation corresponding to the stochastic differential equation (1.1), then time series analysis for the difference can be investigated. From the view of time series analysis, it is natural to consider the model based on dependent data. For the model, the difference should be given by a sequence of dependent random variables.

    In [7], Yoshihara studied a stochastic differential equation such that

    X(tk)X(tk1)=(ν+σξk1)X(tk1),

    where {ξk} is a strictly stationary stochastic sequence. This difference equation corresponds to a Black-Scholes type stochastic differential equation. Under suitable conditions on {ξk}, Yoshihara showed an almost sure convergence theorem for the solution of the difference equation above by using results related to strong Wiener approximations of partial sums of some dependent random variables. Yoshihara's result was extended to multi-dimensional cases in [4], and application to finance models were also considered in [5] and [6]. Following the previous studies, we consider strong approximation of linear stochastic differential equations by weakly dependent random variables in this paper.

    Let {ξk}={(ξk,1,,ξk,d)} be a strictly stationary sequence of d-dimensional centered random vectors defined on a complete probability space (Ω,F,P) and satisfies the strong mixing condition

    α(n)=supAM0BMn|P(AB)P(A)P(B)|0(n), (1.2)

    where Mba(a<b) denotes the σ-algebra generated by {ξa,,ξb}. Under some conditions on {ξk}, which shall be mentioned below, there exists a d×d matrix Γ=(γq,q)q,q=1,,d such that

    {γq,q= Eξ21,q+2i=2Eξ1,qξi,q,γq,q= Eξ1,qξ1,q+i=2(Eξ1,qξi,q+ξ1,qξi,q). (1.3)

    We write

    R=(rq,q)q,q=1,,d=(Eξ1,qξ1,q)q,q=1,,d.

    We remark that if {ξk} is a sequence of independent and identically distributed Rd-valued random variables, then Γ equals to R. In the paper, we always assume that d×d matrix Γ and R are always positive definite. Let {W(t),t0}={(W1(t),,Wd(t)),t0} be a d-dimensional Wiener process with covariance matrix Γ, i.e.,

    EW(t)W(t)T=tΓforall t0,

    where WT is the transpose of the matrix W and set {Ft,t0} as a filtration generated by the Wiener process.

    For a stationary sequence {ξk}, the following strong invariance principle was shown by Liu and Lin [2]. We remark that their result holds under more general conditions:

    Proposition 1.1. Let {ξk} be a stationary strong mixing sequence of centered d-dimensional random vectors. Assume that

    (i) E[ξk]=0,

    (ii) ξk has the third moment,

    (iii) for some τ>0 there exists positive constant c such that α(n)cn3(1+τ).

    Then, on a richer probability space we can redefine the sequence {ξk} together with a d-dimensional Brownian motion {W(t)} whose covariance matrix is Γ such that

    |ktξkW(t)|=O(t1/4) (1.5)

    almost surely, where O(t) is the Landau notation, that is, O(t) is any R-valued sequence for which lim supt(O(t)/t)<.

    Afterward, for the strong approximation (1.5) of Proposition 1.1 we use a notation such that

    ktξkW(t)a.s.

    We denote by AnA a.s. the almost surely convergence of random variables.


    2. Setting and result

    Let A(t),B1(t),,Bd(t) be p×p-matrix functions and a(t),b1(t),,bd(t) be p-dimensional vector functions. We assume that all components of the above matrix and vector functions have continuous bounded derivatives on [0,). On (Ω,F,P), we denote a d-dimensional Wiener process with covariance matrix Γ by {˜W(t),t0}={(˜W1(t),,˜Wd(t)),t0}. It is known that the d-dimensional linear stochastic differential equation

    dX(t)={A(t)X(t)+a(t)}dt+dq=1{Bq(t)X(t)+bq(t)}d˜Wq(t),X(0)=xRp (2.1)

    has a unique solution X(t), which is given by

    Φ(t){x+t0Φ(s)1(a(s)dq=1Bq(s)bq(s))ds+dq=1toΦ(s)1bq(s)d˜Wq(s)}, (2.2)

    where Φ(t) denotes the solution of the matrix stochastic differential equation such that

    dΦ(t)=A(t)Φ(t)dt+dq=1Bq(t)Φ(t)d˜Wq(t),Φ(0)=I, (2.3)

    where I denote the identity matrix.

    We remark that unlike the scalar homogeneous linear equation, we cannot solve (2.3) explicitly for its fundamental solution Φ(t), even when all of the matrices are constant. If A,B1,,Bd are constant matrices and commutative, then we obtain the following explicit expression for the fundamental matrix solution:

    ˜Φ(t)=exp{(A12Rdq=1B2q)tdq=1Bq˜Wq(t)}.

    See Section 2.2 in [1] for more detail. Hence, approximation of solution of stochastic differential equation is a critical tool.

    We fix T>0 arbitrarily. For an arbitrary positive integer n, let m=[n1/6] and l=[n/m], where [x] is a floor function. Using them, we define time point for a sufficiently large n such that

    {si,j= (il+jlm)T, i=0,,l1, j=1,m,s0,0= 0. (2.4)

    We identify sk,0 with sk1,m for k=1,,l.

    Corresponding to (2.1), we consider the following difference equation:

    ΔX(n)(si,j)=X(n)(si,j)X(n)(si,j1):={A(si,j1)X(n)(si,j1)+a(si,j1)}Tlm+dq=1{Bq(si,j1)X(n)(si,j1)+bq(si,j1)}ξim+j,qTlm,X(n)(0)=x Rp. (2.5)

    In addition, corresponding to (2.3), we also consider the following difference equation:

    ΔΦ(n)(si,j)=Φ(n)(si,j)Φ(n)(si,j1):={A(si,j1)Tlm+dq=1Bq(si,j1)ξim+j,qTlm}Φ(n)(si,j1),Φ(n)(0)=I. (2.6)

    For the stochastic difference equation (2.6), we obtain a convergence result by using (i) of Theorem 1 and Theorem 2 in [4].

    Proposition 2.1. We assume that a strictly stationary strong mixing sequence {ξ} satisfies that (i) E[|ξi|6+δ1]< with δ1>0 and (ii) α(n)cn3(1+τ) with some τ(δ)=τ>0. Then, we have the following:

    (1) The solution of (2.6) (we denote by Φ(n)(T)) is given by

    l1i=0mj=1exp{I+A(si,j1)Tlm+dq=1Bq(si,j1)ξim+j,qTlm}.

    (2) On a richer probability space we can redefine the sequence {ξk} together with a d-dimensional Brownian motion {W(t)} whose covariance matrix is Γ such that Φ(n)(T) converges almost surely to a matrix Φ(T) which corresponds to the solution of the stochastic differential equation (2.3).

    Using the solutions Φ(n)(T) and Φ(T), we obtain the solution of the stochastic difference (2.5) as follows:

    Theorem 2.1. Under the same assumptions as those in Proposition 2.1, we obtain the following:

    (1) The solution of (2.5) (we denote by X(n)(T)) is given by

    Φ(n)(T)[x+l1i=0mj=1Φ(n)(si,j)1{a(si,j)+dq,q=1Bq(si,j1)bq(si,j1)ξim+j,qξim+j,q}Tlm+l1i=0mj=1dq=1Φ(n)(si,j1)1bq(si,j1)ξim+j,qTlm].

    (2) On a richer probability space we can redefine the sequence {ξk} together with a p-dimensional Brownian motion {W(t)} whose covariance matrix is Γ such that X(n)(T) converges almost surely to a matrix X(T) which corresponds to the solution of the stochastic differential equation (2.2).


    3. Proof

    Assertion (1) is shown by constructing X(n) in assertion (2). We consider the following difference equation:

    ΔZ(n)(si,j)=Z(n)(si,j)Z(n)(si,j1):=Φ(n)(si,j1)1{a(si,j1)dq,q=1Bq(si,j1)bq(si,j1)ξim+j,qξim+j,q}Tlm+Φ(n)(si,j1)1dq=1bq(si,j1)ξim+j,qTlm,Z(0)=x.

    Then for each (si,j)0il1,1jm we have that

    Δ(Φ(n)Z(n))=ΔΦ(n) Z(n)+Φ(n) ΔZ(n)+ΔΦ(n) ΔZ(n)=:U(n)1+U(n)2+U(n)3. (3.1)

    Firstly, we put that X(t):=Φ(t)Z(t). Then, Theorem 1 in [4] implies that

    l1i=0mj=1U(n)1=l1i=0mj=1{A(si,j1)Tlm+dq=1Bq(si,j1)ξim+j,qTlm}Φ(n)(si,j1)Z(n)(si,j1)T0A(s)X(s)ds+dq=1T0Bq(s)X(s)dWq(s)a.s. (3.2)

    Secondly, we consider that

    l1i=0mj=1U(n)2=l1i=0mj=1a(si,j)Tlml1i=0mj=1dq,q=1Bq(si,j1)bq(si,j1)ξim+j,qξim+j,qTlm+l1i=0mj=1dq=1bq(si,j1)ξim+j,qTlm=:V(n)1V(n)2+V(n)3. (3.3)

    Since a(t) has continuous bounded derivatives on [0,), it is obvious that

    V(n)1T0a(s)ds. (3.4)

    For V(n)2, we use the law of large numbers for {ξk}. Then, we obtain that

    V(n)2dq,q=1rq,qT0Bq(s)bq(s)dsa.s. (3.5)

    Further, using Theorem 1 in [4] again, we have that

    V(n)3dq=1T0bq(s)dWq(s)a.s. (3.6)

    Finally, we consider that

    l1i=0mj=1U(n)3=l1i=0mj=1{A(si,j1)Tlm+dq=1Bq(si,j1)ξim+j,qTlm}[{a(si,j1)dq,q=1Bq(si,j1)bq(si,j1)ξim+j,qξim+j,q}Tlm+dq=1bq(si,j1)ξim+j,qTlm]. (3.7)

    Since under the condition of the theorem on {ξk}, there exists a covariant matrix Γ. We thus obtain that the right hand side of (3.7) equals almost surely to

    Tlml1i=0mj=1dq,q=1Bq(si,j1)bq(si,j1)ξm(i1)+j,qξm(i1)+j,q+O((lm)3/2). (3.8)

    To estimate the first term of the right hand side of (3.8), we consider the term of random variables. Then we obtain that

    1lml1i=0mj=1ξm(i1)+j,qξm(i1)+j,qrq,qa.s. (3.9)

    In addition, all components of Bq and bq(1qp) have continuous bounded derivatives. Thus, there exists some c>0 such that for any 1jm

    Bq(si,j1)bq(si,j1)Bq(si,0)bq(si,0)j1k=1Bq(si,k)bq(si,k)Bq(si,0)bq(si,0)c(j1)lmcl,

    which implies that

    Tlml1i=0mj=1Bq(si,j1)bq(si,j1)ξm(i1)+j,qξm(i1)+j,qrq,qT0Bq(s)bq(s)dsa.s. (3.10)

    Hence, by (3.9) and (3.10) we obtain that

    l1i=0mj=1ΔΦ(n)(si,j)ΔZ(n)(si,j)pq,q=1rq,qT0Bq(s)bq(s)dsa.s. (3.11)

    Combining (3.1), (3.2), (3.3)-(3.6), (3.7) and (3.11), we obtain that

    l1i=0mj=1Δ(Φ(n)(si,j)Z(n)(si,j))T0A(s)X(s)ds+pq=1Bq(s)X(s)dWq(s)+T0a(s)ds+pq=1T0bq(s)dWq(s)a.s.=T0(A(s)X(s)+a(s))ds+pq=1T0(Bq(s)X(s)+bq(s))dWq(s)a.s.

    On the other hand, the solution X(n)(T) of the difference equation

    ΔX(n)(si,j)=X(n)(si,j)X(n)(si,j1):=Δ(Φ(n)(si,j)Z(n)(si,j)),X(n)(0)=x

    satisfies that

    X(n)(T)=x+l1i=0mj=1ΔX(n)(si,j).

    Thus, we obtain that

    X(n)(T)x+T0(A(s)X(s)+a(s))ds+pq=1T0(Bq(s)X(s)+bq)dWq(s)a.s,

    which shows assertion (2) of Theorem 2.1.


    Acknowledgments

    The authors thank Professor R. Naz, Professor M. Torrisi, Professor Ⅰ. Naeem and Professor C.W. Soh for giving an opportunity to talk in AIMS 2016 Meeting, Orlando, Florida, USA. The research is supported by JSPS KAKENHI Grant number 26800063.

    This paper had been almost finished before the second author, Ken-ichi Yoshihara, passed away. I express my gratitude to him for constant encouragement. I will cherish the memory of him forever. I thank the referee for his/her comments which help improve this paper.


    Conflict of Interest

    All authors declare no conflicts of interest in this paper.


    [1] P. E. Kloden and E. Platen, Numerical solution of stochastic differential equations, Springer-Verlag, Berlin Heidelberg, 1992.
    [2] W. Liu and Z. Lin, Strong approximation for a class of stationary processes, Stoch. Proc. Appl., 119 (2009), 249-280.
    [3] X. Mao, Stochastic differential equations and applications, second edition. Horwood Publishing Limited, Chichester, 2008.
    [4] H.Takahashi, S. Kanagawa and K. Yoshihara, Asymptotic behavior of solutions of some difference equations defined by weakly dependent random vectors, Stoch. Anal. Appl., 33 (2015), 740-755.
    [5] H. Takahashi, T. Saigo, S. Kanagawa and K. Yoshihara, Optimal portfolios based on weakly dependent data, Dyn. Syst. Differ. Equ. Appl. , AIMS Proceedings, (2015), 1041-1049.
    [6] H.Takahashi, T. Saigo and K. Yoshihara, Approximation of optimal prices when basic data are weakly dependent, Dyn. Contin. Discrete Impuls. Syst. Ser. B, 23 (2016), 217-230.
    [7] K. Yoshihara, Asymptotic behavior of solutions of Black-Scholes type equations based on weakly dependent random variables, Yokohama Math. J., 58 (2012), 1-15.
  • This article has been cited by:

    1. Yousef Alnafisah, A new order from the combination of exact coupling and the Euler scheme, 2022, 7, 2473-6988, 6356, 10.3934/math.2022353
    2. Yousef Alnafisah, A New Approach to Compare the Strong Convergence of the Milstein Scheme with the Approximate Coupling Method, 2022, 6, 2504-3110, 339, 10.3390/fractalfract6060339
    3. Masataka Hashimoto, Hiroshi Takahashi, On the rate of convergence of Euler–Maruyama approximate solutions of stochastic differential equations with multiple delays and their confidence interval estimations, 2023, 8, 2473-6988, 13747, 10.3934/math.2023698
  • Reader Comments
  • © 2017 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(5130) PDF downloads(1032) Cited by(3)

Article outline

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog