Message Board

Dear readers, authors and reviewers,you can add a message on this page. We will reply to you as soon as possible!

2021 Volume 46 Issue 3
Article Contents

BO En-peng, XIONG Xiang-tuan. A Fractional Tikhonov Method for Sideways Fractional Heat Equation[J]. Journal of Southwest China Normal University(Natural Science Edition), 2021, 46(3): 119-125. doi: 10.13718/j.cnki.xsxb.2021.03.018
Citation: BO En-peng, XIONG Xiang-tuan. A Fractional Tikhonov Method for Sideways Fractional Heat Equation[J]. Journal of Southwest China Normal University(Natural Science Edition), 2021, 46(3): 119-125. doi: 10.13718/j.cnki.xsxb.2021.03.018

A Fractional Tikhonov Method for Sideways Fractional Heat Equation

More Information
  • Corresponding author: XIONG Xiang-tuan
  • Received Date: 07/04/2020
    Available Online: 20/03/2021
  • MSC: O241.1

  • We consider the sideways fractional heat equation in the quarter plane. A new fractional Tikhonovmethod is given to solve the problem. It overcomes the effect of over-smoothing of classical Tikhonov method. We give the selection of prior and posterior regularization parameters of the new method. It makes the error estimation between the exact solution and the approximate solution of the problem reach Hölder optimally.
  • 加载中
  • [1] SEIDMAN T I, ELDEN L. An 'Optimal Filtering' Method for the Sideways Heat Equation[J]. Inverse Problems, 1990, 6(4): 681-696. doi: 10.1088/0266-5611/6/4/013

    CrossRef Google Scholar

    [2] XIONG X T, FU C L, LI H F. Fourier Regularization Method of a Sideways Heat Equation for Determining Surface Heat Flux[J]. Journal of Mathematical Analysis and Applications, 2006, 317(1): 331-348. doi: 10.1016/j.jmaa.2005.12.010

    CrossRef Google Scholar

    [3] DORROH J R, RU X P. The Application of the Method of Quasi-Reversibility to the Sideways Heat Equation[J]. Journal of Mathematical Analysis and Applications, 1999, 236(2): 503-519. doi: 10.1006/jmaa.1999.6462

    CrossRef Google Scholar

    [4] TAUTENHAHN U. Optimality for Ill-Posed Problems under General Source Conditions[J]. Numerical Functional Analysis and Optimization, 1998, 19(3-4): 377-398. doi: 10.1080/01630569808816834

    CrossRef Google Scholar

    [5] CHEN W, YE L J, SUN H G. Fractional Diffusion Equations by the Kansa Method[J]. Computers & Mathematics WithApplications, 2010, 59(5): 1614-1620.

    Google Scholar

    [6] GORENFLO R, MAINARDI F, MORETTI D, et al. Discrete Random Walk Models for Space-Time Fractional Diffusion[J]. Chemical Physics, 2002, 284(1-2): 521-541. doi: 10.1016/S0301-0104(02)00714-0

    CrossRef Google Scholar

    [7] ZASLAVSKY G M. Chaos, Fractional Kinetics, and Anomalous Transport[J]. Physics Reports, 2002, 371(6): 461-580. doi: 10.1016/S0370-1573(02)00331-9

    CrossRef Google Scholar

    [8] SCALAS E, GORENFLO R, MAINARDI F. Fractional Calculus and Continuous-Time Finance[J]. Physica A: Statistical Mechanics and Its Applications, 2000, 284(1-4): 376-384. doi: 10.1016/S0378-4371(00)00255-7

    CrossRef Google Scholar

    [9] 赵梅妹. 改进的分数阶辅助方程方法及其在非线性空间-时间分数阶微分方程中的应用[J]. 西南师范大学学报(自然科学版), 2018, 43(11): 24-29.

    Google Scholar

    [10] LIU F, ANH V, TURNER I. Numerical Solution of the Space Fractional Fokker-Planck Equation[J]. Journal of Computational and Applied Mathematics, 2004, 166(1): 209-219. doi: 10.1016/j.cam.2003.09.028

    CrossRef Google Scholar

    [11] OZALPN, MIZRAK O O. Fractional Laplace Transform Method in the Framework of the CTIT Transformation[J]. Journal of Computational and Applied Mathematics, 2017, 317: 90-99. doi: 10.1016/j.cam.2016.11.039

    CrossRef Google Scholar

    [12] VILELA MENDES R. A Fractional Calculus Interpretation of the Fractional Volatility Model[J]. Nonlinear Dynamics, 2008, 55(4): 395-399.

    Google Scholar

    [13] LI M, XIONG X T. On a Fractional Backward Heat Conduction Problem: Application to Deblurring[J]. Computers & Mathematics With Applications, 2012, 64(8): 2594-2602.

    Google Scholar

    [14] QIAN Z, FENG X L. A Fractional Tikhonov Method for Solving a Cauchy Problem of Helmholtz Equation[J]. Applicable Analysis, 2017, 96(10): 1656-1668. doi: 10.1080/00036811.2016.1254776

    CrossRef Google Scholar

    [15] 薛雪敏, 熊向团, 庄娥, 等. 时间分数阶反扩散问题的一种新的分数次Tikhonov方法[J]. 高校应用数学学报A辑, 2018, 33(4): 441-452.

    Google Scholar

    [16] PODLUBNY I. Fractional Differential Equations[M]. San Diego: Academic Press, 1999.

    Google Scholar

    [17] ELDÉN L, BERNTSSON F, REGINSKA T. Wavelet and Fourier Methods for Solving the Sideways Heat Equation[J]. SIAM Journal on Scientific Computing, 2000, 21(6): 2187-2205. doi: 10.1137/S1064827597331394

    CrossRef Google Scholar

    [18] FENG X L, FU C L, CHENG H. A Regularization Method for Solving the Cauchy Problem for the Helmholtz Equation[J]. Applied Mathematical Modelling, 2011, 35(7): 3301-3315. doi: 10.1016/j.apm.2011.01.021

    CrossRef Google Scholar

    [19] XIONG X T. A Regularization Method for a Cauchy Problem of the Helmholtz Equation[J]. Journal of Computational and Applied Mathematics, 2010, 233(8): 1723-1732. doi: 10.1016/j.cam.2009.09.001

    CrossRef Google Scholar

    [20] XIONG X T, XUE X M. Fractional Tikhonov Method for an Inverse Time-Fractional Diffusion Problem in 2-Dimensional Space[J]. Bulletin of the Malaysian Mathematical Sciences Society, 2020, 43(1): 25-38. doi: 10.1007/s40840-018-0662-5

    CrossRef Google Scholar

    [21] 熊向团. 抛物型偏微分方程中几类反问题的正则化理论及算法[D]. 兰州: 兰州大学, 2007.

    Google Scholar

    [22] QIAN Z, FU C L. Regularization Strategies for a Two-Dimensional Inverse Heat Conduction Problem[J]. Inverse Problems, 2007, 23(3): 1053-1068. doi: 10.1088/0266-5611/23/3/013

    CrossRef Google Scholar

  • 加载中
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

Article Metrics

Article views(1304) PDF downloads(107) Cited by(0)

Access History

Other Articles By Authors

A Fractional Tikhonov Method for Sideways Fractional Heat Equation

    Corresponding author: XIONG Xiang-tuan

Abstract: We consider the sideways fractional heat equation in the quarter plane. A new fractional Tikhonovmethod is given to solve the problem. It overcomes the effect of over-smoothing of classical Tikhonov method. We give the selection of prior and posterior regularization parameters of the new method. It makes the error estimation between the exact solution and the approximate solution of the problem reach Hölder optimally.

  • 在一些工程问题中,人们往往需要确定一个物体的表面温度,但又无法在物体表面直接测量,就必须由物体内部某固定位置的温度来反演表面温度,这就是所谓的逆热传导问题,或称为热传导方程的侧边值问题. 这类问题是严重不适定的,许多学者给出了不同的正则化方法,包括最优滤波法[1]、Fourier截断法[2]、拟逆法[3]、谱方法[4]等. 近年来,分数阶微分问题广泛应用于数学界和工程界[5-12], 用分数阶微分方程去阐释热传导过程是一种较为成功的途径. 针对分数阶热传导方程的侧边值问题,本文给出一种分数次Tikhonov正则化方法,此方法既包含了经典Tikhonov方法的优点,又克服了经典Tikhonov解的过度光滑性,而且在做误差估计时也体现了该方法的简洁性. 这种分数次Tikhonnov方法是文献[13]提出的一种新的正则化方法,文献[14]应用这种方法讨论了Helmholtz方程的Cauchy问题,文献[15]讨论时间分数阶反扩散问题时也用到此方法. 本文的方法和结果为解决更复杂的问题奠定了基础.

1.   问题不适定性
  • 考虑以下不适定问题:

    其中时间分数阶导数 $\frac{\partial^{\alpha} u}{\partial t^{\alpha}}$由文献[16]中α(0 < α < 1)阶Caputo导数定义.

    注1  当α=1时,问题(1)是经典的热传导方程的侧边值问题[17].

    实际问题中,我们只能考虑给定在x=x处的温度测量数据gδ(t), 并由它来反演问题(1)在区间(0, x)上的解u(x, t), 其中gδ(t)∈L2($\mathbb{R}$)满足

    这里常数δ > 0表示测量误差,‖·‖表示L2-范数.

    进一步,我们假设先验界如下:

    其中N是给定的正数.

    为了在频域中考虑问题(1), 我们将关于t的函数延拓到整个实轴,令t < 0的部分为零. 定义函数f(t)的Fourier变换如下:

    相应的函数 $\hat{f}(\xi)$的Fourier逆变换为:

    对问题(1)关于变量t作Fourier变换:

    得到问题(8)的解为

    其中

    η的实部和虚部分别表示为:

    因此

    根据(9)式得问题(1)的精确解为:

    由Parseval等式[18]和条件(5)得:

    注意到当|ξ|→∞时$\mathrm{e}^{\bar{x}} \sqrt{(\mathrm{i} \xi)^{\alpha}} \rightarrow \infty$ 式中精确数据g(t)的Fourier变换 $\hat{g}(\xi)$必须是快速衰减的,但测量数据gδ(ξ)属于L2($\mathbb{R}$), 其高频部分衰减的速率不会有这么快,这样g(t)的很小扰动就会导致与(9)式对应的 $\hat{u}^{\delta}(x, \xi)$不存在Fourier逆变换. 为了得到问题(1)的稳定近似解,下面将给出一种分数次Tikhonov正则化方法.

2.   先验情况下正则化参数的选择
  • 我们考虑用分数次Tikhonov正则化方法来解决不适定问题(1). 由(9)式可知

    因此, $\hat{u}(x, \xi)$可写为

    在频域空间构造分数次Tikhonov正则化解为:

    其中:μ > 0是正则化参数,σ是分数次参数.

    接下来为了使证明简便,给出一个辅助引理.

    引理1[19]  如果常数μ > 0, 0 < p < q, 对于变量s≥0, 有如下不等式成立:

    定理1  设u(x, t)由(13)式给出,且条件(4)和(5)成立,那么当μ被选取为

    成立如下误差估计式:

    并称uμδ(x, t)为问题(1)的正则化解.

      由三角不等式和Parseval等式可得

    $I_{1}=\left\|\hat{u}_{\mu}^{\delta}(x, \bullet)-\hat{u}_{\mu}(x, \bullet)\right\|$, $I_{2}=\| \hat{u}_{\mu}(x, ・) -\hat{u}(x, ・) \|$ , 下面分别估计I1, I2,

    由(15), (16)式可得:

    因此,

    由(19)式可得

    下面考虑I2的估计,并应用先验界(5)得

    由(23), (24)式可知

    根据(19)式得

    故由(20), (25), (26)式得到如下Hölder型最优误差估计式

3.   后验情况下正则化参数的选择
  • 在这一部分将研究后验情况下正则化参数的选择,并且通过Morozov's偏差原理[20]确定一个正则化参数μ满足如下方程:

    其中: $\frac{1}{2} \leqslant \sigma \leqslant 1$是常数,μ是正则化参数.

    下面的结论是明显的.

    引理2  设 $\beta(\mu)=\left\|\frac{1}{1+\mu|T(0, \xi)|^{2 \sigma}} \hat{g}^{\delta}(\xi)-\hat{g}^{\delta}(\xi)\right\|$, 并且 $0<\tau \delta<\left\|\hat{g}^{\delta}(\xi)\right\|$, 则

    (a) β(μ)为连续函数;

    (b) $\lim\limits_{\mu \rightarrow 0} \beta(\mu)=0$;

    (c) $\lim\limits_{\mu \rightarrow 0} \beta(\mu)=\left\|\hat{g}^{\delta}(\xi)\right\|_{L^{2}(\mathbb{R})} ;$

    (d) β(μ)为严格单调增函数.

    注2  根据引理2可知,若 $0<\tau \delta<\left\|\hat{g}^{\delta}(\xi)\right\|$, 则方程(27)的解存在且唯一.

    引理3  若μ满足方程(27), 则有以下不等式成立:

      由方程(27)和三角不等式得

    引理4  若μ满足方程(27), 则得到以下不等式:

    根据(19)式可得

    定理2  设条件(4)和先验界(5)成立,且正则化参数μ是通过Morozov's偏差原理(27)确定,则有如下误差估计成立:

    其中 $c=\left(\frac{\tau}{\tau-1}\right)^{1-\frac{x}{\bar{x}}}(\tau+1)^{\frac{x}{\bar{x}}}$.

      令 $I^{2}=\| u(x, \bullet)-u_{\mu}^{\delta}(x$, - $) \|^{2}$, 由Parseval等式及(28), (29)式可知

    由Hölder不等式及(23), (24)式可得:

    由(19)式可得:

    根据引理4可得

    因此

    $c=\left(\frac{\tau}{\tau-1}\right)^{1-\frac{x}{\bar {x}}}(\tau+1)^{\frac{x}{\bar {x}}}$

    注3  注意到这里仅考虑问题(1)在区间(0, x)上的精确解与正则解之间的误差估计,而未涉及边界x=0的情况,但是,我们通过(21)和(32)式看出,当x=0时,误差估计式不收敛. 若给出先验界‖u(0, ·)‖pN, 其中p > 0, N > 0, ‖·‖p表示Hp-范数,则由文献[21-22]知收敛速率呈对数型.

4.   总结
  • 不适定问题的种类有很多,需要我们构建正则化方法来解决此问题,每种正则化方法都有各自的优缺点. 本文使用一种新的分数次Tikhonv正则化方法求解分数阶热传导方程的侧边值问题,通过先验和后验的正则化参数的选择得到了Hölder型的误差估计,结果证实了该方法的简洁性和有效性. 这种新的分数次Tikhonov方法也可能适用于Laplace方程的柯西问题等其他不适定问题,这有待于进一步研究.

Reference (22)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return