邓飞其 莫浩艺
摘要
本文回顾了近年来随机微分方程数值方法的稳定性的研究成果.作为相关话题,收敛性问题也有所涉猎.以经典It型随机微分方程、
中立型随机泛函微分方程、Markov跳随机微分方程和Poisson跳随机微分方程为代表,主要介绍了几类数值方法稳定性研究的成果.
这些方法包括常见的 EulerMaruyama 方法、Backward EulerMaruyama方法、θ方法、分步方法等.文中
分析了关于稳定性等价性定理经典论文的学术思路,提出了随机微分方程数值计算与仿真所面临的挑战及所要解决的问题.关键词随机微分方程;数值格式;稳定性;仿真
中图分类号P393
文献标志码A
1华南理工大学自动化科学与工程学院,广州,510640
2广东工业大学应用数学学院,广州,510006
1典型数值方法及其收敛性
由于大多数随机微分方程解析解的显式表达式都很难得到,快速高效的数值算法对于随机微分方程的应用
显得格外重要.对于随机微分方程数值解的研究,大体来说可以分为两类:有限时间的收敛性和随着时间变量
趋于无穷的渐近性.本节主要是对有限时间的收敛性的相关研究进行回顾.
其中向量n=Yn+f(Yn)Δ+g(Yn)Δ.通过重复运用Taylor展式,对方程f和g展开的阶数越高,所获得格式的收敛阶数会越高,可以达到15阶或20阶,但其形式也更复杂,从而影响其广泛应用.请参见专著[1].
针对不同模型和精度,格式的构造和分析有许多后续进展,取得了丰富的成果,这里不一一列举.例如,Liang等[4]研究了一类线性随机Volterra积分方程,在Lipschitz条件下,证明了EM方法是1阶强收敛的;Wang等[5]分析了带有加性噪声的半线性随机偏微分方程隐式Euler方法的弱收敛性.
在众多数值算法中,EM型算法由于结构简单、易于编程等特点受到很多学者的关注[69].它是所有随机微分方程数值算法里最简单的一种.经典的Euler型算法,即EM算法,是常微分方程的向前Euler算法的自然推广.上面提到,在全局Lipschitz条件下,经典的EM方法是强05阶收敛的,但是当漂移项或扩散项不满足全局Lipschitz条件时,EM方法将不收敛.Hutzenthaler等[10]对于这种不收敛性(发散性)给出了严格的证明.
那么,针对非Lipschitz方程,各类格式是否也可用?精度又如何?为此,学者们开展了系列的探讨.例如,为了处理一类漂移项不满足全局Lipschitz条件的随机微分方程,特别是当漂移项满足多项式增长时,Hutzenthaler等[11]提出了具有05阶强收敛性的Tamed(驯服) Euler方法.简单来说,Tamed Euler方法在经典的EM算法基础上增加了对漂移项的控制,它的格式如下:
同时,该文还利用类似的技巧提出了具有1阶强收敛性的Balanced Milstein方法.
我们注意到,上述不同种类Euler型算法虽然结构不同,但是证明思路多是先证明数值解和解析解的p阶矩有界,然后再根据不同的算法格式证明强收敛性和收敛阶.这种证明思路或多或少借鉴了文献[7].在文献[7]中,作者给出了在已知Euler型数值解矩有界时推导收敛性和收敛速率的技巧.
另一种利用数值解局部收敛性推导全局收敛性的技巧也非常重要.在全局Lipschitz条件以及Khasminskii条件下,数值解的全局误差可以由局部误差推导出的结论,可分别在文献[20]和文献[21]中找到.
综上所述,当我们想构造显式的Euler型方法来数值逼近漂移项和扩散项不满足全局Lipschitz条件的随机微分方程时,采用的方法主要是在经典的EM方法基础上利用一些约束方式来控制漂移项和扩散项.一个很自然的问题是:以上这些理论上具有05阶(或者1阶)强收敛性的方法孰优孰劣?对于这个开放性问题,也许文献[22]中关于最优强收敛系数K1的讨论是一个思路.
2数值方法的稳定性
我们先来谈谈微分方程数值计算格式的稳定性的来源.
微分方程的数值计算格式稳定性概念的提出源于计算数学领域对数值计算舍入误差传播问题的考虑.
众所周知,由于计算工具限制等各种原因,在数值计算过程中,舍入误差在所难免,某一步计算的舍入误差一定会随计算格式带入往后各步,也就是说,舍入误差将向后传播.
如果计算格式对该误差具有敏感性,则该误差将随格式进行传播,被累计、被放大,甚至产生蝴蝶效应.当年,费肯鲍姆就是因为运用数字计算格式时出现了初值误差而发现了混沌现象.如果格式对该误差不敏感,则该误差的影响将被逐渐消除,无积累效应,不被严重放大.即在一定条件下得到控制,从而被最终屏蔽.基于此考虑,在计算数学领域提出了微分方程数值计算格式的稳定性概念,用以描述計算格式对舍入误差的敏感性.如果一个格式对舍入误差敏感,
则称格式不稳定;否则,称其稳定.所以,微分方程数值计算格式的稳定性,是一个定性概念.
微分方程数值计算格式稳定,意味着计算格式可以自行消化舍入误差,不在传播中因累计而放大.
最常见的数值计算格式稳定性概念是绝对稳定性,在此不赘述.
本文所述计算格式稳定性概念与此相同.在系统与控制科学领域,我们同样需要考虑格式的收敛性(逼近度)和稳定性.我们的目的是:如何将计算格式用于系统仿真,并通过系统仿真分析(原)系统的稳定性.
在随机系统数值计算方面,数值方法的收敛性和稳定性是学者们主要讨论的两大类内容.由于大部分随机系统的非线性性和耦合性,很难求出其解析解.所以通过离散化的数值方法来研究系统的稳定性是一种有效的途径,它是窥探系统内部结构和性态的一种手段.目前探讨的问题主要是:
1)在一定条件下,比照连续模型与离散格式的稳定性,看看离散格式是否复制了连续模型的稳定性质;
2)连续模型与离散格式的稳定性的逻辑互推.
本文将主要讨论几类It随机微分方程数值方法的稳定性.数值方法的稳定性主要包括:矩意义下的渐近稳定、p阶矩指数稳定、几乎必然指数稳定、依概率稳定、A稳定等[2]
.其研究内容和方法要比确定型常微分系统丰富很多.下面,先介绍本文讨论的几大类稳定性定义,其中p>0.
值得指出,对连续模型解的稳定性也有类似于上面的定义,只需要在上述定义中将数值解Xk换成解析解x(t),k→∞替换为t→∞即可.在这些稳定性定义中,p阶矩指数稳定可推出渐近稳定,而在文献中,一般同时关注几乎必然指数稳定与矩指数稳定性,但实际上它们之间并无必然联系,因此,都是分开单独推证得出相关结论.如果附加一定的条件,比如线性增长条件,则由p阶矩指数稳定可推出几乎必然指数稳定[2].一般而言,p阶矩指数稳定可以通过估计解的矩E|x(t,x0)|p来得到,这时需要借助某个适当的Lyapunov函数V(t,x)来估计EV(t,x(t)),因此Lyapunov方法是研究矩稳定的一个很有效的方法.与矩指数稳定性不同,几乎必然指数稳定是一种轨道稳定,它依赖于解的轨道估计,通常有下面三种方法可推出幾乎必然指数稳定:1)由解的矩指数稳定,利用Chebyshev不等式推出解的几乎必然指数稳定;2)利用非负半鞅收敛定理,直接证明解的几乎必然指数稳定;3)通过指数鞅不等式和BorelCantelli引理证明解的几乎必然指数稳定.
文献中,对随机微分方程数值解稳定性的研究,一般采用直接的推证方法,很少套用Lyapunov稳定性定理,但其中
同样含有Lyapunov函数或者泛函的思想方法.
下面,从模型推广与方法创新的角度,分别介绍几类It型随机微分方程数值方法稳定性研究所取得的进展.
21中立型随机泛函微分方程
经典的It随机微分方程(SDE)已经被许多学者研究[24,2630].随着科学技术的高速发展,实践中的许多领域,如生物工程、机械工程等都涉及到时间滞后的现象.由于时滞带的存在,系统状态的变化不仅与当前的时间状态相关,而且还与过去的历史状态有关.从而,诞生了描述这类系统的随机时滞微分方程:
中提出,其意义是将确定中立型泛函微分方程推广到随机中立型泛函微分方程.后来,Mao[3233]分别讨论了中立型泛函型随机微分方程解析解的均方指数稳定性以及运用
Razumikhin技术证明解的指数稳定性.其后,相关学者开展一系列出色的研究工作,如Liao等[34]研究了中立型随机时滞微分方程解析解的几乎必然指数稳定性;Luo等[35]为了克服文献[36]中要求函数满足线性增长条件和
时滞为常数,提出局部Lipschitz条件,建立了相应的稳定性定理,证明了中立型时滞微分方程解析解的指数均方稳定性;
如果随机θ方法满足假设1和假设2.那么,研究θ方法的p阶矩指数稳定性可以得到方程(27)解的p阶矩指数稳定性.这类结果揭示了:数值格式的稳定性与连续模型稳定性在逻辑上可以互推.因此,这是目前数值研究中不多见的一种研究思路,
其进一步的研究,也相当具有挑战性.
4对逼近度方法学术思路的分析
从终极目标看,我们的研究目的是提供可靠的理论保障,使我们能从系统仿真结果推断系统的渐近性态,如稳定性.因此,需要先确定系统解析解与数值解稳定性可以从逻辑上互推的性质,提供严密的理论依据.显然,为实现这类互推,需要建立两种解之间的关联,否则,不可能存在互推.而这种关联,用逼近度描述正好合适,其原因在于:1)我们设计方程求解的数值格式,分析其逼近度是最主要的一项基础工作,对我们的需要来说,是顺手的事;2)符合互推稳定性的需要.所以,在毛学荣教授及其合作者的系列论文中,提出了这类假设,即数值格式具有高于零阶的逼近度,其实就是局部截断误差、收敛性[9596].当然,我们也注意到,这类假设直接涉及方程的解析解和数值解本身,而问题是:我们并不具体知道它们.正是因为方程难以求解,我们才借助数值计算与仿真.所以,其实这类条件本身是不能直接验证的.因此,需要采用其他条件对此予以保证,例如Lipschitz条件.在Mao[98]提出一般理论之前,以前的相关文献直接采用Lipschitz条件,高于零阶的逼近度是其自然推论.从这个角度来看,采用Lipschitz条件而不是采用逼近度的假设,更加符合研究结果的描述与验证.但是,如果有Lipschitz条件,则当然有了高于零阶的逼近度,所以,逼近度条件其实更弱.这里,为清晰和比较起见,我们不妨称逼近度方法所得结果为命题,而采用Lipschitz条件的结果为判据.
5随机微分方程数值计算与仿真所面临的的挑战
51关于等价性结论与数值仿真结果的意义与运用
通过数值仿真真的可以确定系统解析解的稳定性吗? 难!
实际上,当我们在一定条件下建立了解析解与数值解之间的稳定性等价性定理,我们所得的是系统稳定性之间的等价性,是系统与系统之间的互推关系,是集合与集合之间的互推关系,而不是两个系统个别解之间的互推关系.原理上,我们的仿真一次只确定一个解的渐近性态,而一般地,基于一个解的渐近性态,例如就是指数渐近稳定性,我们还是不足以推断整个数值格式的稳定性,更不能推断关于解析解的任何性质,哪怕我们就是想推断一个解的性质,那也不能,因为没有依据.那么,我们如何从仿真结果确定数值格式以及原系统解析解系统的稳定性呢? 首先,我们需要有等价性结论作基础; 其次,我们需要确证数值格式稳定.在假设第一个问题已有结论的前提下,我们来看第二个问题,即确定数值格式稳定的难度.为讨论方便,我们先放下随机微分方程,回到确定型方程.简单说,这个问题其实就是差分格式通解的构成问题.如果差分格式的通解可以由若干互不相关的特解构成,例如就是线性组合,而我们又能确定若干互不相关的特解的渐近性态,那问题就解决了.所以,如果我们的格式是n阶常系数线性差分格式,则需要n个互不相关的特解的渐近性态,也就是说:我们需要n个初值线性无关的特解的仿真结果.当然,如果n=1,一个仿真结果就够了.但是,如果方程再略微复杂,则难以有如此明确的结论,问题的难度也陡增.例如,如果我们的格式是非线性格式、随机格式,因为一般不存在关于通解构成的基础理论,我们就不完全知道需要用多少个特解来确定通解(即便是存在所谓的通解).因此,也就不知道需要用多少个仿真来确定格式的稳定性.我们认为:可以用多少个、用什么样的仿真结果确定数值格式的稳定性从而可以推断原系统解析解的稳定性是一个具有挑战性的问题.
52面向渐近稳定性的等价性结论
因为推导的需要,目前的等价性结论都是面向指数稳定性的.但是,实际上,数字计算与仿真提供的是具有直观属性的数字与图形.一般地,从一個仿真结果很难看出一个数值解是否真的就是指数稳定,只能看出是否是渐近稳定.只有面向渐近稳定性的等价性结论才有实用价值.因此,我们需要建立面向渐近稳定性的等价性结论,而这,其论证难度陡增,也是今后可以考虑但具有相当难度的一个挑战课题.
结束语与致谢:
由于时间、篇幅和水平所限,本文所综述的工作只是相关工作中的一点点,难免挂一漏万,敬请谅解.
在本文写作过程中,吴付科教授、宋明辉教授、宗小峰博士、刘暐博士、付余老师、杨慧子博士及赵桂华老师等给予了大力指导、支持与协助.在此,向为本文写作给予了支持的所有师生表示衷心的感谢.
参考文献
References
[1]Kloeden P E,Platen E.Numerical solution of stochastic differential equations[M].Springer Verlag Berlin,Germany Google Scholar,1992
[2]Mao X.Stochastic differential equations and applications[M].Elsevier,2007
[3]Kloeden P E,Platen E.Higherorder implicit strong numerical schemes for stochastic differential equations[J].Journal of Statistical Physics,1992,66(1/2):283314
[4]Liang H,Yang Z,Gao J.Strong superconvergence of the EulerMaruyama method for linear stochastic Volterra integral equations[J].Journal of Computational and Applied Mathematics,2017,317:447457
[5]Wang X,Gan S.Weak convergence analysis of the linear implicit euler method for semilinear stochastic partial differential equations with additive noise[J].Journal of Mathematical Analysis and Applications,2013,398(1):151169
[6]Higham D J.Stochastic ordinary differential equations in applied and computational mathematics[J].IMA Journal of Applied Mathematics,2011,76:449474
[7]Higham D J,Mao X,Stuart A M.Strong convergence of Eulertype methods for nonlinear stochastic differential equations[J].SIAM Journal on Numerical Analysis,2002,40(3):10411063
[8]Yu Z.Almost sure and mean square exponential stability of numerical solutions for neutral stochastic functional differential equations[J].International Journal of Computer Mathematics,2015,92(1):132150
[9]Pang S,Deng F,Mao X.Almost sure and moment exponential stability of EulerMaruyama discretizations for hybrid stochastic differential equations[J].Journal of Computational and Applied Mathematics,2008,213(1):127141
[10]Hutzenthaler M,Jentzen A,Kloeden P E.Strong and weak divergence in finite time of Eulers method for stochastic differential equations with nonglobally lipschitz continuous coefficients[J].Proceedings of the Royal Society A Mathematical,Physical and Engineering Sciences,2009,467(2130):15631576
[11]Hutzenthaler M,Jentzen A,Kloeden P E.Strong convergence of an explicit numerical method for SDEs with nonglobally lipschitz continuous coefficients[J].The Annals of Applied Probability,2012:16111641
[12]Hutzenthaler M,Jentzen A.Numerical approximations of stochastic differential equations with nonglobally Lipschitz continuous coefficients[M].American Mathematical Society,2015
[13]Wang X,Gan S.The tamed Milstein method for commutative stochastic differential equations with nonglobally Lipschitz continuous coefficients[J].Journal of Difference Equations and Applications,2013,19(3):466490
[14]Zong X,Wu F,Huang C.Convergence and stability of the semitamed Euler scheme for stochastic differential equations with nonLipschitz continuous coefficients[J].Applied Mathematics and Computation,2014,228:240250
[15]Mao X.The truncated EulerMaruyama method for stochastic differential equations[J].Journal of Computational and Applied Mathematics,2015,290:370384
[16]Mao X.Convergence rates of the truncated EulerMaruyama method for stochastic differential equations[J].Journal of Computational and Applied Mathematics,2016,296:362375
[17]Guo Q,Liu W,Mao X,et al.The partially trumcated EulerMaruyama method and its stability and bounded ness[J].Applied Numerical Mathematics,2017,115:235251
[18]Guo Q,Liu W,Mao X,et al.The truncated milstein method for stochastic differential equations[J].arXiv:1704.04135[math.NA],2017
[19]Zhang Z,Ma H.Orderpreserving strong schemes for SDEs with locally Lipschitz coefficients[J].Applied Numerical Mathematics,2017,112:116
[20]Milstein G N,Tretyakov M V.Stochastic numerics for mathematical physics[M].Springer Science & Business Media,2013
[21]Tretyakov M V,Zhang Z.A fundamental meansquare convergence theorem for SDEs with locally Lipschitz coefficients and its applications[J].SIAM Journal on Numerical Analysis,2013,51(6):31353162
[22]MüllerGronbach T.The optimal uniform approximation of systems of stochastic differential equations[J].The Annals of Applied Probability,2002,12(2):664690
[23]Wang W,Chen Y.Meansquare stability of semiimplicit Euler method for nonlinear neutral stochastic delay differential equations[J].Applied Numerical Mathematics,2011,61(5):696701
[24]Zong X,Wu F,Choice of θ and meansquare exponential stability in the stochastic theta method of stochastic differential equations[J].Journal of Computational and Applied Mathematics,2014,255:837847
[25]Zhou S,Xie S,Fang Z.Almost sure exponential stability of the backward EulerMaruyama discretization for highly nonlinear stochastic functional differential equation[J].Applied Mathematics and Computation,2014,236:150160
[26]Zong X,Wu F,Huang C.Preserving exponential mean square stability and decay rates in two classes of theta approximations of stochastic differential equations[J].Journal of Difference Equations and Applications,2014,20(7):10911111
[27]Huang C.Exponential mean square stability of numerical methods for systems of stochastic differential equations[J].Journal of Computational and Applied Mathematics,2012,236(16):40164026
[28]王小捷.随机微分方程数值算法研究[D].长沙:中南大学,2012
WANG Xiaojie.Numerical study of stochastic differential equations[D].Changsha:Central South University,2012
[29]李燕.随机系统的稳定性与数值策略的研究[D].武汉:华中科技大学,2014
LI Yan.Research on stability and numerical strategy of stochastic systems[D].Wuhan:Huazhong University of Science and Technology,2014
[30]王冠军.几类随机非线性系统的动力学分析[D].南京:东南大学,2009
WANG Guanjun.Dynamical analysis for several classes of stochastic nonlinear systems[D].Nanjing:Southeast University,2016
[31]Kolmanovsky V,Nosov V.Stability of neutraltype functional differential equations[J].Nonlinear Analysis:Theory,Methods & Applications,1982,6(9):873910
[32]Mao X.Exponential stability in mean square of neutral stochastic differential functional equations[J].Systems & Control Letters,1995,26(4):245251
[33]Mao X.Razumikhintype theorems on exponential stability of neutral stochastic differential equations[J].SIAM Journal on Mathematical Analysis,1997,28(2):389401
[34]Liao X,Mao X.Almost sure exponential stability of neutral stochastic differential difference equations[J].Journal of Mathematical Analysis and Applications,1997,212(2):554570
[35]Luo Q,Mao X,Shen Y.New criteria on exponential stability of neutral stochastic differential delay equations[J].Systems & Control Letters,2006,55(10):826834
[36]Mao X.Asymptotic properties of neutral stochastic differential delay equations[J].Stochastics:An International Journal of Probability and Stochastic Processes,2000,68(3/4):273295
[37]胡榮.中立型随机泛函微分方程解的渐近性质[D].武汉:华中科技大学,2009
HU Rong.Asymptotic properties of solutions of neutral stochastic functional differential equations[D].Wuhan:Huazhong University of Science and Technology,2009
[38]Huang C.Mean square stability and dissipativity of two classes of theta methods for systems of stochastic delay differential equations[J].Journal of Computational and Applied Mathematics,2014,259:7786
[39]Liu L,Zhu Q.Mean square stability of two classes of theta method for neutral stochastic differential delay equations[J].Journal of Computational and Applied Mathematics,2016,305:5567
[40]Zong X,Wu F,Huang C.Exponential mean square stability of the theta approximations for neutral stochastic differential delay equations[J].Journal of Computational and Applied Mathematics,2015,286:172185
[41]Zhou S.Exponential stability of numerical solution to neutral stochastic functional differential equation[J].Applied Mathematics and Computation,2015,266:441461
[42]Zhou S,Hu C.Numerical approximation of stochastic differential delay equation with coefficients of polynomial growth[J].Calcolo,2016:122
[43]李啟勇.几类随机延迟微分方程数值方法的稳定性分析[D].长沙:中南大学,2012
LI Qiyong.Stability analysis of numerical methods for several classes of stochastic delay differential equations[D].Changsha:Central South Unirersity,2012
[44]马强.几类随机微分方程的保结构数值方法[D].哈尔滨:哈尔滨工业大学,2012
MA Qiang.Structurepreserving numerical methods for several classes of stochastic differential equations[D].Harbin:Harbin Institute of Technology,2012
[45]吴瑞华.几类随机种群模型渐近性质的研究[D].哈尔滨:哈尔滨工业大学,2014
WU Ruihua.Research on asymptotic properties of several stochastic population systems[D].Harbin:Harbin Institute of Technology,2014
[46]Yu Z,Liu M.Almost surely asymptotic stability of numerical solutions for neutral stochastic delay differential equations[J].Discrete Dynamics in Nature and Society,2011,Doi:10.1155/2011/217672
[47]Yu Z.The improved stability analysis of the backward Euler method for neutral stochastic delay differential equations[J].International Journal of Computer Mathematics,2013,90(7):14891494
[48]Zong X,Wu F.Exponential stability of the exact and numerical solutions for neutral stochastic delay differential equations[J].Applied Mathematical Modelling,2016,40(1):1930
[49]Wu F,Mao X.Numerical solutions of neutral stochastic functional differential equations[J].SIAM Journal on Numerical Analysis,2008,46(4):18211841
[50]Wu F,Hu S,Huang C.Robustness of general decay stability of nonlinear neutral stochastic functional differential equations with infinite delay[J].Systems & Control Letters,2010,59(3):195202
[51]Jankovic S,Jovanovic M.The pth moment exponential stability of neutral stochastic functional differential equations[J].Filomat,2006,20(1):5972
[52]Kats I I,Krasovskii N.On the stability of systems with random parameters[J].Journal of Applied Mathematics and Mechanics,1960,24(5):12251246
[53]Krasovskii N,Lidskii E.Analytical design of controllers in systems with random attributes[J].Automation and Remote Control,1961,22(1/2/3):10211025
[54]Milshtein G N,Repin Y M.The action of a Markov process on a system of differential equations[J].Differentsialnye Uravneniya,1969,5(8):13711384
[55]Yuan C,Mao X.Convergence of the EulerMaruyama method for stochastic differential equations with Markovian switching[J].Mathematics and Computers in Simulation,2004,64(2):223235
[56]Wang L,Xue H.Convergence of numerical solutions to stochastic differential delay equations with Poisson jump and Markovian switching[J].Applied Mathematics and Computation,2007,188(2):11611172
[57]王振東.基于可控马尔科夫链的跳跃系统控制问题研究[D].合肥:中国科学技术大学 ,2014
WANG Zhendong.Control of Markovian jump systems with controllable Markov chain[D].Hefei:University of Science and Technology of China,2014
[58]Li R,Meng H,Dai Y.Convergence of numerical solutions to stochastic delay differential equations with jumps[J].Applied Mathematics and Computation,2006,172(1):584602
[59]Mao X,Yuan C,Yin G.Numerical method for stationary distribution of stochastic differential equations with Markovian switching[J].Journal of Computational and Applied Mathematics,2005,174(1):127
[60]Mao X,Shen Y,Yuan C.Almost surely asymptotic stability of neutral stochastic differential delay equations with Markovian switching[J].Stochastic Processes and their Applications,2008,118(8):13851406
[61]Zhou S,Wu F.Convergence of numerical solutions to neutral stochastic delay differential equations with Markovian switching[J].Journal of Computational and Applied Mathematics,2009,229(1):8596
[62]Mao X.Stability of stochastic differential equations with Markovian switching[J].Stochastic Processes and Their Applications,1999,79(1):4567
[63]Mao X,Yuan C.Stochastic differential equations with Markovian switching[M].World Scientific,2006
[64]Higham D J,Mao X,Yuan C.Preserving exponential meansquare stability in the simulation of hybrid stochastic differential equations[J].Numerische Mathematik,2007,108(2):295325
[65]Mao X,Shen Y,Gray A.Almost sure exponential stability of backward EulerMaruyama discretizations for hybrid stochastic differential equations[J].Journal of Computational and Applied Mathematics,2011,235(5):12131226
[66]Mao X,Szpruch L.Strong convergence rates for backward EulerMaruyama method for nonlinear dissipativetype stochastic differential equations with superlinear diffusion coeffcients[J].Stochastics:An International Journal of Probability and Stochastic Processes,2013,85(1):144171
[67]Higham D J,Kloeden P E.Strong convergence rates for backward Euler on a class of nonlinear jumpdiffusion problems[J].Journal of Computational and Applied Mathematics,2007,205(2):949956
[68]Wang X,Gan S.The improved splitstep backward Euler method for stochastic differential delay equations[J].International Journal of Computer Mathematics,2011,88(11):23592378
[69]Mao X,Yuan C,Yin G.Approximations of EulerMaruyama type for stochastic differential equations with Markovian switching,under nonlipschitz conditions[J].Journal of Computational and Applied Mathematics,2007,205(2):936948
[70]Li R,Meng H,Chang Q.Exponential stability of numerical solutions to SDDEs with Markovian switching[J].Applied Mathematics and Computation,2006,174(2):13021313
[71]Mao X,Matasov A,Piunovskiy A B,et al.Stochastic differential delay equations with Markovian switching[J].Bernoulli,2000,6(1):7390
[72]Higham D J,Kloeden P E.Numerical methods for nonlinear stochastic differential equations with jumps[J].Numerische Mathematik,2005,101(1):101119
[73]Higham D J,Kloeden P E.Convergence and stability of implicit methods for jumpdiffusion systems[J].International Journal of Numerical Analysis and Modeling,2006,3(2):125140
[74]Song M,Yang H,Liu M,et al.Strong convergence of the tamed Euler method for stochastic differential equations with piecewise continuous arguments and Poisson jumps under nonglobally Lipschitz continuous coefficients[J].Filomat,accepted
[75]Wang L,Mei C,Xue H.The semiimplicit Euler method for stochastic differential delay equation with jumps[J].Applied Mathematics and Computation,2007,192(2):567578
[76]Chalmers G D,Higham D J.Convergence and stability analysis for implicit simulations of stochastic differential equations with random jump magnitudes[J].Discrete and Continuous Dynamical Systems Series B,2008,9(1):4764
[77]趙桂华.几类带跳随机微分方程数值解的收敛性和稳定性[D].哈尔滨:哈尔滨工业大学,2009
ZHAO Guihua.Convergence and stability of numerical solutions for several classes of stochastic differential equations with jumps[D].Harbin:Harbin Institute of Technology,2009
[78]Fu Y,Zhao W,Zhou T.Multistep schemes for forward backward stochastic differential equations with jumps[J].Journal of Scientific Computing,2016,69(2):651672
[79]Wang X,Gan S.Compensated stochastic theta methods for stochastic differential equations with jumps[J].Applied Numerical Mathematics,2010,60(9):877887
[80]Hu L,Gan S.Convergence and stability of the balanced methods for stochastic differential equations with jumps[J].International Journal of Computer Mathematics,2011,88(10):2089 2108
[81]Li Q,Gan S.Almost sure exponential stability of numerical solutions for stochastic delay differential equations with jumps[J].Journal of Applied Mathematics and Computing,2011,37(1/2):541557
[82]Wu F,Mao X,Szpruch L.Almost sure exponential stability of numerical solutions for stochastic delay differential equations[J].Numerische Mathematik,2010,115(4):681697
[83]Tan J,Wang H.Meansquare stability of the EulerMaruyama method for stochastic differential delay equations with jumps[J].International Journal of Computer Mathematics,2011,88(2):421429
[84]Li Q,Gan S,Wang X.Compensated stochastic theta methods for stochastic differential delay equations with jumps[J].International Journal of Computer Mathematics,2013,90(5):10571071
[85]Jacob N,Wang Y,Yuan C.Stochastic differential delay equations with jumps,under nonlinear growth condition[J].Stochastics:An International Journal of Probability and Stochastics Processes,2009,81(6):571588
[86]Jacob N,Wang Y,Yuan C.Numerical solutions of stochastic differential delay equations with jumps[J].Stochastic Analysis and Applications,2009,27(4):825853
[87]Tan J,Wang H,Guo Y.Existence and uniqueness of solutions to neutral stochastic functional differential equations with Poisson jumps[J].Abstract and Applied Analysis,2012(3):509512
[88]Liu D,Yang G,Zhang W.The stability of neutral stochastic delay differential equations with Poisson jumps by fixed points[J].Journal of Computational and Applied Mathematics,2011,235(10):31153120
[89]Tan J,Wang H,Guo Y,et al.Numerical solutions to neutral stochastic delay differential equations with Poisson jumps under local Lipschitz condition[J].Mathematical Problems in Engineering,2014(1):111
[90]胡琳.几类带泊松跳随机微分方程数值方法的收敛性与稳定性[D].长沙:中南大学,2012
HU Lin.Convergence and stability of numerical methods several classes of stochastic differential equations with Poissondriven jumps[D].Changsha:Central South University,2012
[91]Mo H,Zhao X,Deng F.Exponential meansquare stability of the θmethod for neutral stochastic delay differential equations with jumps[J].International Journal of Systems Science,2017,48(3):462470
[92]Song M,Yang H,Liu M.Convergence and stability of impulsive stochastic differential equations[J].International Journal of Computer Mathematics,2016:19
[93]宗小峰.隨机微分方程的数值分析及随机稳定化[D].武汉:华中科技大学,2014
ZONG Xiaofeng.Numerical analysis of stochastic differential equations and stochastic stabilization[D].Wuhan:Huazhong University of Science and Technology,2014
[94]Fu Y,Zhao W.An explicit secondorder numerical scheme to solve decoupled forward backward stochastic equations[J].East Asian Journal on Applied Mathematics,2014,4(4):368385
[95]Higham D J,Mao X,Stuart A M.Exponential meansquare stability of numerical solutions to stochastic differential equations[J].LMS Journal of Computation and Mathematics,2003,6:297313
[96]Mao X.,Exponential stability of equidistant EulerMaruyama approximations of stochastic differential delay equations[J].Journal of Computational and Applied Mathematics,2007,200(1):297316
[97]Miloevic′ M.The EulerMaruyama approximation of solutions to stochastic differential equations with piecewise constant arguments[J].Journal of Computational and Applied Mathematics,2016,298:112
[98]Mao X.Almost sure exponential stability in the numerical simulation of stochastic differential equations[J].SIAM Journal on Numerical Analysis,2015,53(1):370389
Abstract
In this paper,a survey is given for the investigation on the stability of numerical schemes of stochastic differential equations in the past years.As a related topic,the convergence of the schemes is involved.The paper introduces the achieved results by literatures for the classical It stochastic differential equations,stochastic functional differential equations of the neutral type, and the stochastic differential equations with Markov or Poisson jumps.The involved numerical schemes include the EulerMaruyama scheme,the Backward EulerMaruyama scheme,the θ scheme,and the splitstep scheme,etc.The paper analyzes the academic thoughts in some classical literatures on the stability equivalence theorems and proposes some problems and challenges for further investigations on the numerical computations and simulations of stochastic differential equations at the end of the paper.
Key wordsstochastic differential equations;numerical schemes;stability;simulations