带Poisson跳的随机时滞微分方程的矩有界性 Moment Boundedness of Stochastic Delay Differential Equations with Poisson Jumps

doi:10.12677/PM.2021.118170

Pure Mathematics
Vol. 11 No. 08 ( 2021 ), Article ID: 44657 , 6 pages
10.12677/PM.2021.118170

带Poisson跳的随机时滞微分方程的矩有界性

林宇璇，李光洁^*

●How to Cite this Article

广东外语外贸大学数学与统计学院，广东广州

收稿日期：2021年7月9日；录用日期：2021年8月12日；发布日期：2021年8月19日

摘要

研究一类带Poisson跳的随机时滞微分方程解的有界性。运用随机分析以及不等式技巧证明了该方程的解是 $p (p \in (0, 1))$ -阶矩有界的。

关键词

随机时滞微分方程，矩有界性，Poisson跳

Moment Boundedness of Stochastic Delay Differential Equations with Poisson Jumps

Yuxuan Lin, Guangjie Li^*

School of Mathematics and Statistics, Guangdong University of Foreign Studies, Guangzhou Guangdong

Received: Jul. 9^th, 2021; accepted: Aug. 12^th, 2021; published: Aug. 19^th, 2021

ABSTRACT

The boundedness of solutions for stochastic delay differential equations with Poisson jumps is investigated in this paper. By using stochastic analysis and inequality techniques, it is proved that the p-th $p (p \in (0, 1))$ moment of the solution of such equations is bounded.

Keywords:Stochastic Delay Differential Equations, Moment Boundedness, Poisson Jumps

This work is licensed under the Creative Commons Attribution International License (CC BY 4.0).

http://creativecommons.org/licenses/by/4.0/

1. 引言

随机微分方程广泛地应用于生物学、数学、经济、金融领域以及控制工程等诸多领域，见文献 [1] [2] [3] 及其中的参考文献。在许多实际工程系统中，时滞是这些实际系统经常发生振荡、不稳定性和性能差的原因。因此，很有必要研究带有时滞的随机微分方程，学者们在研究随机时滞微分方程方面做出了许多努力，参阅文献 [4] [5] [6] [7]。王等人在文献 [8] 和文献 [9] 中分别研究了带分布时滞的线性随机时滞微分方程的矩有界性和线性随机时滞微分方程的二阶矩有界性。文献 [10] 研究了脉冲随机时滞微分方程的有界性。

Brown运动是连续的随机过程，可以用来刻画连续的随机系统。然而许多实际的随机系统会遭遇跳跃形式的随机突发扰动。在这些情形下，不能用Brown运动来刻画这些随机系统，因此将带Poisson跳的过程引入到这些随机系统中来处理这些实际情形是合理的 [11] [12] [13]。本文研究了一类带Poisson跳的随机时滞微分方程的 $p (p \in (0, 1))$ -阶矩有界性。该文结构如下：第2节给出预备知识，第3节给出本文的主要结果。

2. 预备知识

记 $R^{n}$ 表示n-维的欧式空间。若 $x \in R^{n}$ ， $| x |$ 表示欧式范数。记 $R = (- \infty, + \infty)$ ， $R^{+} = [0, + \infty)$ 。记一个完备的概率空间为 $(Ω, F, P)$ ，其滤子 $F_{t}$ 满足通常的条件，即 ${F_{t}}_{t \geq 0}$ 是右连续的且 $F_{0}$ 包含所有的零测集。令 $δ > 0$ ， $C ([- δ, 0]; R^{n})$ 表示所有定义在 $[- δ, 0]$ 上的 $R^{n}$ -值连续函数 $φ$ 。 $C_{F_{0}}^{b} (Ω; R^{n})$ 表示所有 $F_{0}$ -可测的且定义在 $C ([- δ, 0]; R^{n})$ 上的有界函数全体，且其上的范数为 $‖ φ ‖ = \sup_{- δ \leq θ \leq 0} | φ (θ) |$ 。对 $\forall x, y \in R^{n}$ ， $〈 x, y 〉$ 和 $x^{T} y$ 均表示内积。

${r (t)}_{t \geq 0}$ 表示的是定义在完备概率空间上的一个取值为 $S = {1, 2, \dots, N}$ 的右连续时齐Markov链，其生成元(密度矩阵) $Γ = {(γ_{i j})}_{N \times N}$ 由转移概率矩阵确定，即

$P {r (t + Δ) = j | r (t) = i} = {\begin{array}{l} γ_{i j} Δ + o (Δ), & 当 i \neq j, \\ 1 + γ_{i i} Δ + o (Δ), & 当 i = j, \end{array}$

这里， $Δ > 0$ ， $\lim_{Δ \to 0} o (Δ) / Δ = 0$ ，对于 $i \neq j$ ， $γ_{i j} \geq 0$ 表示从状态i到状态j的转移速率且 $γ_{i i} = - \sum_{j \neq i} γ_{i j}$ 。作为一个事实， $r (\cdot)$ 的每个样本轨道是一个右连续的阶梯函数，且在 $R^{+} = [0, + \infty)$ 上的任何一个有限区间里至多存在有限个跳跃点(参阅 [14] )。

考虑如下形式的带Poisson跳的随机时滞微分方程：

$\begin{array}{l} d x (t) = [f (x (t), t, r (t)) + u (x (t - δ), t, r (t))] d t + g (x (t), t, r (t)) d ω (t) \\ + h (x (t), t, r (t)) d N (t), t \geq t_{0}, \end{array}$ (1)

初始值 $x_{t_{0}} = φ (θ) = {x (t_{0} + θ) : - δ \leq θ \leq 0} \in C_{F_{0}}^{b} (Ω; R^{n})$ ， $r (0) = i_{0} \in S$ ，其中 $f, g, h, u : R^{n} \times R^{+} \times S \to R^{n}$ ， $ω (t)$ 和 $N (t)$ 分别是定义在该概率空间上的1-维Brown运动和强度为 $λ$ 的Poisson过程，这里假设Markov链 $r (t)$ 与Brown运动 $B (t)$ 和Poisson过程 $N (t)$ 是相互独立的。 $\tilde{N} (t) = N (t) - λ t$ 表示 $N (t)$ 的补偿Poisson过程。本文假设对于任意的 $x_{1}, x_{2} \in R^{n}, t \in R^{+}$ 和 $i \in S$ ，函数 $f, g, h, u$ 满足如下条件：

$\begin{array}{l} | f (x_{1}, t, i) - f (x_{2}, t, i) | \leq L_{1} | x_{1} - x_{2} |, \\ | g (x_{1}, t, i) - g (x_{2}, t, i) | \leq L_{2} | x_{1} - x_{2} |, \\ | h (x_{1}, t, i) - h (x_{2}, t, i) | \leq L_{3} | x_{1} - x_{2} |, \\ | u (x_{1}, t, i) - u (x_{2}, t, i) | \leq K | x_{1} - x_{2} | . \end{array}$ (2)

且对 $\forall t \geq 0$ 和 $i \in S$ ，假设 $f (0, t, i) = g (0, t, i) = h (0, t, i) = 0$ 。因此，当方程(1)的初始值 $x_{0} = 0$ 时，该方程存在平凡解 $x (t) \equiv 0$ 。进一步，经计算对 $\forall (x, i, t) \in R^{n} \times S \times R^{+}$ 有

$| f (x, t, i) | \leq L_{1} | x |, | g (x, t, i) | \leq L_{2} | x |, | h (x, t, i) | \leq L_{3} | x |, | u (x, t, i) | \leq K | x | .$

注在假设条件(2)下，方程(1)存在唯一的全局解且 $E {| x (t) |}^{2} < \infty$ ( $\forall t \geq 0$ )。

令 $C^{2, 1} (R^{n} \times R^{+} \times S; R^{+})$ 表示的是对 $\forall i \in S$ 关于变量x二阶连续可导且关于变量t一阶连续可导的全体非负函数 $V (x, t, i)$ 的集合。给定任意的 $V (x, t, i) \in C^{2, 1} (R^{n} \times R^{+} \times S; R^{+})$ ，定义算子

$\begin{matrix} L V (x, t, i) = V_{t} (x, t, i) + V_{x} (x, t, i) [f (x, t, i) + u (x, t, i)] \\ + \frac{1}{2} t r a c e [g^{T} (x, t, i) V_{x x} (x, t, i) g (x, t, i)] \\ + λ [V (x + h (x, t, i), t, i) - V (x, t, i)] \\ + \sum_{j = 1}^{N} γ_{i j} V (x, t, j), \end{matrix}$

其中， $V_{t} (x, t, i) = \frac{\partial v (x, t, i)}{\partial t}$ ， $V_{x x} (x, t, i) = {(\frac{\partial^{2} v (x, t, i)}{\partial x_{i} \partial x_{j}})}_{n \times n}$ ，

$V_{x} (x, t, i) = (\frac{\partial v (x, t, i)}{\partial x_{1}}, \frac{\partial v (x, t, i)}{\partial x_{2}}, \dots, \frac{\partial v (x, t, i)}{\partial x_{n}}) .$

3. 主要结果

令 $p \in (0, 1)$ 。本节将证明方程(1)的解是p-阶矩有界的。

定理若条件(1)成立，则对所有的 $t_{0} \geq 0$ 和 $T \geq 0$ ，

$\sup_{t_{0} \leq t \leq t_{0} + T + δ} E {| x (t) |}^{p} \leq M_{1}^{\frac{p}{2}} E {‖ φ ‖}^{2},$ (3)

$E (\sup_{t_{0} \leq t \leq t_{0} + T + δ} {| x (t) |}^{p}) \leq M_{2}^{\frac{p}{2}} E {‖ φ ‖}^{2},$ (4)

其中， $M_{1} = M_{1} (p, δ, T) = (1 + k δ) e^{[2 (L_{1} + λ L_{3}) + L_{2}^{2} + λ L_{3}^{2} + 2 K] (T + δ)}$ ， $M_{2} = M_{2} (p, δ, T) = 5 [1 + L_{1}^{2} (T + δ) + K^{2} (T + δ) + 4 L_{2}^{2} + 2 λ^{2} L_{3}^{2} (T + δ) + 8 λ L_{3}^{2}] (T + δ) M_{1}$ 。

证明(1) 对 ${| x (t) |}^{2}$ 应用Itô公式，由方程(1)可得

$\begin{matrix} d {| x (t) |}^{2} = 2 〈 x (t), f (x (t), t, r (t)) + u (x (t - δ), t, r (t)) 〉 d t \\ + 2 〈 x (t), g (x (t), t, r (t)) 〉 d ω (t) + {| g (x (t), t, r (t)) |}^{2} d t \\ + [{| x (t) + h (x (t), t, r (t)) |}^{2} - x^{2}] d N ( t ) \end{matrix}$

$\begin{array}{l} = 2 〈 x (t), f (x (t), t, r (t)) + u (x (t - δ), t, r (t)) 〉 d t \\ + 2 〈 x (t), g (x (t), t, r (t)) 〉 d ω (t) + {| g (x (t), t, r (t)) |}^{2} \\ + λ [h^{2} (x (t), t, r (t)) + 2 x^{T} (t) h (x (t), t, r (t))] d t \\ + [h^{2} (x (t), t, r (t)) + 2 x^{T} (t) h (x (t), t, r (t))] d \tilde{N} (t) \end{array}$ (5)

对(3)式两边同时取期望，计算得

$\begin{matrix} E {| x (t) |}^{2} = E {| x (t_{0}) |}^{2} + 2 E \int_{t_{0}}^{t} x^{T} (s) f (x (s), s, r (s)) d s \\ + 2 E \int_{t_{0}}^{t} x^{T} (s) u (x (s - δ), s, r (s)) d s \\ + 2 E \int_{t_{0}}^{t} {| g (x (s), s, r (s)) |}^{2} d s \\ + λ E \int_{t_{0}}^{t} [h^{2} (x (s), s, r (s)) + 2 x^{T} (s) h (x (s), s, r (s))] d s . \end{matrix}$ (6)

利用Hölder不等式可得

$\begin{array}{l} 2 K E \int_{t_{0}}^{t} | x (s) | | x (s - δ) | d s \\ \leq K E \int_{t_{0}}^{t} {| x (s) |}^{2} d s + K E \int_{t_{0}}^{t} {| x (s - δ) |}^{2} d s \\ = K E \int_{t_{0}}^{t} {| x (s) |}^{2} d s + K E \int_{t_{0} - δ}^{t - δ} {| x (ν) |}^{2} d ν \\ = K E \int_{t_{0}}^{t} {| x (s) |}^{2} d s + K E \int_{t_{0} - δ}^{t_{0}} {| x (ν) |}^{2} d ν + K E \int_{t_{0}}^{t - δ} {| x (ν) |}^{2} d ν \\ \leq 2 K E \int_{t_{0}}^{t} {| x (s) |}^{2} d s + K δ E {‖ φ ‖}^{2} . \end{array}$ (7)

将(7)代入到(6)中可得

$\begin{matrix} E {| x (t) |}^{2} \leq E {| x (t_{0}) |}^{2} + [2 (L_{1} + λ L_{3}) + L_{2}^{2} + λ L_{3}^{2} + 2 K] \int_{t_{0}}^{t} E {| x (s) |}^{2} d s + K δ E {‖ φ ‖}^{2} \\ \leq (1 + K δ) E {‖ φ ‖}^{2} + [2 (L_{1} + λ L_{3}) + L_{2}^{2} + λ L_{3}^{2} + 2 K] \int_{t_{0}}^{t} (\sup_{t_{0} \leq ν \leq s} E {| x (ν) |}^{2}) d s . \end{matrix}$ (8)

从而

$\sup_{t_{0} \leq ν \leq t_{0} + T + δ} E {| x (ν) |}^{2} \leq (1 + K δ) E {‖ φ ‖}^{2} + [2 (L_{1} + λ L_{3}) + L_{2}^{2} + λ L_{3}^{2} + 2 K] \int_{t_{0}}^{t_{0} + T + δ} (\sup_{t_{0} \leq ν \leq s} E {| x (ν) |}^{2}) d s .$ (9)

进一步利用Gronwall不等式得

$\sup_{t_{0} \leq ν \leq t_{0} + T + δ} E {| x (ν) |}^{2} \leq (1 + K δ) e^{[2 (L_{1} + λ L_{3}) + L_{2}^{2} + λ L_{3}^{2} + 2 K] (T + δ)} E {‖ φ ‖}^{2} = M_{1} E {‖ φ ‖}^{2} .$

最后运用Hölder不等式，进一步得

$\sup_{t_{0} \leq t \leq t_{0} + T + δ} E {| x (t) |}^{p} \leq M_{1}^{\frac{p}{2}} E {‖ φ ‖}^{p} .$

(2)由方程(1)可得

$\begin{matrix} x (t) = x (t_{0}) + \int_{t_{0}}^{t} [f (x (s - δ), s, r (s)) + u (x (s - δ), s, r (s))] d s \\ + \int_{t_{0}}^{t} g (x (s), s, r (s)) d ω (s) + \int_{t_{0}}^{t} h (x (s), s, r (s)) d N (s) . \end{matrix}$ (10)

利用Hölder不等式和条件(2)可得

$\begin{matrix} {| x (t) |}^{2} \leq 5 {| x (t_{0}) |}^{2} + 5 {| \int_{t_{0}}^{t} f (x (s), s, r (s)) d s |}^{2} + 5 {| \int_{t_{0}}^{t} u (x (s - δ), s, r (s)) d s |}^{2} \\ + 5 {| \int_{t_{0}}^{t} g (x (s), s, r (s)) d ω (s) |}^{2} + 5 {| \int_{t_{0}}^{t} h (x (s), s, r (s)) d N (s) |}^{2} \\ \leq 5 {| x (t_{0}) |}^{2} + 5 (t - t_{0}) \int_{t_{0}}^{t} f^{2} (x (s), s, r (s)) d s + 5 (t - t_{0}) \int_{t_{0}}^{t} u^{2} (x (s - δ), s, r (s)) d s \\ + 5 {| \int_{t_{0}}^{t} g (x (s), s, r (s)) d ω (s) |}^{2} + 10 λ^{2} (t - t_{0}) \int_{t_{0}}^{t} h^{2} (x (s), s, r (s)) d s \\ + 10 {| \int_{t_{0}}^{t} h (x (s), s, r (s)) d \tilde{N} (s) |}^{2} . \end{matrix}$

利用Doob鞅不等式可得

$\begin{array}{l} E (\sup_{t_{0} \leq t \leq t_{0} + T + δ} {| x (t) |}^{2}) \\ \leq 5 E {| x (t_{0}) |}^{2} + 5 (T + δ) E \int_{t_{0}}^{t_{0} + T + δ} f^{2} (x (s), s, r (s)) d s \\ + 5 (T + δ) E \int_{t_{0}}^{t_{0} + T + δ} u^{2} (x (s - δ), s, r (s)) d s + 20 E \int_{t_{0}}^{t_{0} + T + δ} g^{2} (x (s), s, r (s)) d s \\ + 10 λ^{2} (T + δ) E \int_{t_{0}}^{t_{0} + T + δ} h^{2} (x (s), s, r (s)) d s + 40 λ E \int_{t_{0}}^{t_{0} + T + δ} h^{2} (x (s), s, r (s)) d s \\ \leq 5 E {‖ φ ‖}^{2} + 5 L_{1}^{2} (T + δ) \int_{t_{0}}^{t_{0} + T + δ} E {| x (s) |}^{2} d s + 5 K^{2} (T + δ) \int_{t_{0}}^{t_{0} + T + δ} E {| x (s - δ) |}^{2} d s \\ + 20 L_{2}^{2} \int_{t_{0}}^{t_{0} + T + δ} E {| x (s) |}^{2} d s + 40 λ L_{3}^{2} E \int_{t_{0}}^{t_{0} + T + δ} E {| x (s) |}^{2} d s . \end{array}$

利用(3)，进一步可得

$\begin{array}{l} E (\sup_{t_{0} \leq t \leq t_{0} + T + δ} {| x (t) |}^{2}) \\ \leq 5 [1 + L_{1}^{2} (T + δ) + K^{2} (T + δ) + 4 L_{2}^{2} + 2 λ^{2} L_{3}^{2} (T + δ) + 8 λ L_{3}^{2}] \times (T + δ) M_{1} E {‖ φ ‖}^{2} \\ = : M_{2} E {‖ φ ‖}^{2} . \end{array}$

再次运用不等式可得

$E (\sup_{t_{0} \leq t \leq t_{0} + T + δ} {| x (t) |}^{2}) \leq M_{3}^{\frac{p}{2}} E {‖ φ ‖}^{2} .$

证毕。

基金项目

国家自然科学基金项目(No. 11901398)。

文章引用

林宇璇,李光洁. 带Poisson跳的随机时滞微分方程的矩有界性
Moment Boundedness of Stochastic Delay Differential Equations with Poisson Jumps[J]. 理论数学, 2021, 11(08): 1511-1516. https://doi.org/10.12677/PM.2021.118170

参考文献

1. Mao, X. (1997) Stochastic Differential Equations and Application. Horwood Publication, Chichester.

2. Mao, X. and Yuan, C. (2006) Stochastic Differential Equations with Markovian Switching. Imperial College Press, Lon-don.

3. Arnold, L. (2007) Stochastic Differential Equations: Theory and Applications. World Scientific, Singa-pore.

4. Buckwar, E. (2000) Introduction to the Numerical Analysis of Stochastic Delay Differential Equations. Journal of Computational and Applied Mathematics, 125, 297-307. https://doi.org/10.1016/S0377-0427(00)00475-1

5. Rodkina, A. and Basin, M. (2006) On Delay-Dependent Sta-bility for a Class of Nonlinear Stochastic Delay-Differential Equations. Mathematics of Control, Signals and Systems, 18, 187-197. https://doi.org/10.1007/s00498-006-0163-1

6. Ngoc, P. (2020) New Criteria for Mean Square Exponential Sta-bility of Stochastic Delay Differential Equations. International Journal of Control, 2020, 1-17.

7. Cao, W. and Zhu, Q. (2021) Stability Analysis of Neutral Stochastic Delay Differential Equations via the Vector Lyapunov Function Method. Applied Mathematics and Computation, 405, Article ID: 126257.

8. Wang, Z., Li, X. and Lei, J. (2014) Moment Boundedness of Linear Stochastic Delay Differential Equations with Distributed Delay. Stochastic Process and Application, 124, 586-612. https://doi.org/10.1016/j.spa.2013.09.002

9. Wang, Z., Li, X. and Lei, J. (2017) Second Moment Boundedness of Linear Stochastic Delay Differential Equations. Discrete and Continuous Dynamical Systems, 19, 2963-2991.

10. Xu, L., Dai, Z. and He, D. (2018) Exponential Ultimate Boundedness of Impulsive Sto-chastic Delay Differential Equations. Applied Mathematics Letters, 85, 70-76. https://doi.org/10.1016/j.aml.2018.05.019

11. Wu, S., Dong, H., and Meng, X. (2004) p-Moment Stability of Stochastic Differential Equations with Jumps. Applied Mathematics and Computation, 152, 505-519. https://doi.org/10.1016/S0096-3003(03)00573-3

12. Liu, D., Yang, G. and Zhang, W. (2011) The Stability of Neutral Stochastic Delay Differential Equations with Poisson Jumps by Fixed Points. Journal of Computational and Applied Mathematics, 235, 3115-3120. https://doi.org/10.1016/j.cam.2008.10.030

13. Wei, Y. and Yang, Q. (2018) Dynamics of the Stochastic Low Concentration Trimolecular Oscillatory Chemical System with Jumps.Communications in Nonlinear Science and Nu-merical Simulation, 59, 396-408. https://doi.org/10.1016/j.cnsns.2017.11.019

14. Anderson, W. (1991) Continuous-Time Markov Chains. Springer, Berlin.

NOTES

^*通讯作者。

期刊菜单