American Institute of Mathematical Sciences

December  2020, 7(2): 183-208. doi: 10.3934/jcd.2020008

Solving the inverse problem for an ordinary differential equation using conjugation

 1 Departamento de Ciência da Computação, Universidade Federal do Rio de Janeiro, Caixa Postal 68.530, CEP 21941-590, Rio de Janeiro, RJ, Brazil 2 Departamento de Matemática, Universidade Federal de Juiz de Fora, Juiz de Fora, CEP 36036-900, MG, Brazil 3 Departamento de Matemática y Ciencia de la Computación, Universidad de Santiago de Chile, Casilla 307, Correo 2, Santiago de Chile, Chile 4 Instituto de Matemática Pura e Aplicada, Estrada Dona Castorina 110, CEP 22460-320, Rio de Janeiro, RJ, Brazil

* Corresponding author: Daniel G. Alfaro Vigo

Received  October 2019 Published  July 2020

Fund Project: The second author's work was partially supported by IMPA/CAPES. The third author was partially supported by FAPEMIG under Grant APQ 01377/15 and CNPq under grant 303245/2019-0. The fourth author was partially supported by DICYT grant 041933GM from VRIDEI-USACH

We consider the following inverse problem for an ordinary differential equation (ODE): given a set of data points $P = \{(t_i,x_i),\; i = 1,\dots,N\}$, find an ODE $x^\prime(t) = v (x)$ that admits a solution $x(t)$ such that $x_i \approx x(t_i)$ as closely as possible. The key to the proposed method is to find approximations of the recursive or discrete propagation function $D(x)$ from the given data set. Afterwards, we determine the field $v(x)$, using the conjugate map defined by Schröder's equation and the solution of a related Julia's equation. Moreover, our approach also works for the inverse problems where one has to determine an ODE from multiple sets of data points.

We also study existence, uniqueness, stability and other properties of the recovered field $v(x)$. Finally, we present several numerical methods for the approximation of the field $v(x)$ and provide some illustrative examples of the application of these methods.

Citation: Daniel G. Alfaro Vigo, Amaury C. Álvarez, Grigori Chapiro, Galina C. García, Carlos G. Moreira. Solving the inverse problem for an ordinary differential equation using conjugation. Journal of Computational Dynamics, 2020, 7 (2) : 183-208. doi: 10.3934/jcd.2020008
References:

show all references

References:
Iteration function $D$ (solid line), points of the data set with a noise level $\sigma = 0.5$ (blue points) and the recovered iteration function (red circle) as discussed in Example 1 (Subsection 6.1). Notice that the exact and recovered iteration functions are basically indistinguishable
The exact field $v(x)$ (solid blue line) and the recovered field corresponding to Example 1 (Subsection 6.1), for different values of $\sigma$. For $\sigma = 0.1$ (dashed blue line), $\sigma = 0.5$ (dotted black), $\sigma = 0.9$ (dash-dot green line), $\sigma = 1.5$ (dashed red blue) and $\sigma = 2.9$ (dotted magenta line). Notice that for $\sigma<1$ exact and recovered fields are indistinguishable
Multiple sets of data points (upper plot) and the corresponding synthetic data (lower plot) used in Example 2 (Subsection 6.2). The set 1 (blue points) is used in both cases (a) and (b), whereas the other sets (red circles) are only used in case (b)
Exact and approximate iteration function and its derivative (upper plot) and the associated approximation errors (lower plot) in the interval $(0,1)$, corresponding to case (a) of Example 2 (Subsection 6.2). Notice that the exact functions and their approximations are indistinguishable
Exact and approximated field $v(x)$ (upper plot) and approximation error (lower plot) in the interval $(0,1)$ corresponding to case (a) of Example 2 (Subsection 6.2). Observe that the exact field and its approximation are indistinguishable
Exact and approximated iteration function and its derivative (upper plot) and the associated approximation errors (lower plot) corresponding to case (b) of Example 2 (Subsection 6.2). Notice that the exact functions and their approximations are almost indistinguishable
Exact and approximated field $v(x)$ (upper plot) and approximation error (lower plot) in the interval $(-2.04,2.04)$ corresponding to case (b) of Example 2 (Subsection 6.2). Notice that the exact field and its approximation are indistinguishable
Exact and approximated iteration function and its derivative (upper plot), and approximation errors for the iteration function and its derivative (lower plot); corresponding to Example 3 (Subsection 6.3)
Exact and approximated field (upper plot) and the approximation error (lower plot) corresponding to Example 3 (Subsection 6.3). Notice that the difference between the field and its approximation is only noticeable around $x = 1.2$
 Algorithm 1: Implementation of formula (35) Require: $x_0$, $\epsilon$, functions $D$ and $D^{\prime}$ Ensure: $g(x_0) = q_n$ 1: $x_n=x_0$, error=1, lim=1, $q_n=1$ 2: while error $> \epsilon$ do 3:    last=lim 4:    $q_n=q_n D(x_n)/(x_n D^{\prime}(x_n))$ 5:    $x_n=D(x_n)$ 6:    lim=$q_n$ 7:    error=$|$lim-last$|$/$|$last$|$ 8: end while 9: $q_n = x_o q_n$ 10: return $q_n$
 Algorithm 1: Implementation of formula (35) Require: $x_0$, $\epsilon$, functions $D$ and $D^{\prime}$ Ensure: $g(x_0) = q_n$ 1: $x_n=x_0$, error=1, lim=1, $q_n=1$ 2: while error $> \epsilon$ do 3:    last=lim 4:    $q_n=q_n D(x_n)/(x_n D^{\prime}(x_n))$ 5:    $x_n=D(x_n)$ 6:    lim=$q_n$ 7:    error=$|$lim-last$|$/$|$last$|$ 8: end while 9: $q_n = x_o q_n$ 10: return $q_n$
 Algorithm 2: Implementation of formula (37) Require: $x_0,\dots,x_m$, $\epsilon$, functions $D$ and $D^{\prime}$ Ensure: $g(x_0) = g_0, \dots, g(x_m) = g_m$ 1: $g_0 = \cdots = g_m = 1$, error=1, lim=1 2: $q_0 = D(x_0)/(x_0 D^{\prime}(x_0)), \dots, q_m = D(x_m)/(x_m D^{\prime}(x_m))$ 3: while error $> \epsilon$ do 4:    $gl_0 = g_0, \dots, gl_m = g_m$ 5:    Compute function $g(x)$ interpolating data : $(x_0,g_0), \dots, (x_m,g_m)$ 6:    $g_0= q_0 g(D(x_0)), \dots, g_m= q_m g(D(x_m))$ 7:    error=max($|gl_j-g_j|$)/max($|gl_j|$) 8: end while 9: $g_0 = x_o g_0, \dots, g_m = x_m g_m$ 10: return $g_0, \dots, g_m$
 Algorithm 2: Implementation of formula (37) Require: $x_0,\dots,x_m$, $\epsilon$, functions $D$ and $D^{\prime}$ Ensure: $g(x_0) = g_0, \dots, g(x_m) = g_m$ 1: $g_0 = \cdots = g_m = 1$, error=1, lim=1 2: $q_0 = D(x_0)/(x_0 D^{\prime}(x_0)), \dots, q_m = D(x_m)/(x_m D^{\prime}(x_m))$ 3: while error $> \epsilon$ do 4:    $gl_0 = g_0, \dots, gl_m = g_m$ 5:    Compute function $g(x)$ interpolating data : $(x_0,g_0), \dots, (x_m,g_m)$ 6:    $g_0= q_0 g(D(x_0)), \dots, g_m= q_m g(D(x_m))$ 7:    error=max($|gl_j-g_j|$)/max($|gl_j|$) 8: end while 9: $g_0 = x_o g_0, \dots, g_m = x_m g_m$ 10: return $g_0, \dots, g_m$
Values of the relative errors for the iteration function $D$, its derivative $D'$ and the field $v$, and the stability constant $C_v$ corresponding to example 6.1
 $\sigma$ $\epsilon_{D}$ $\epsilon_{D^{'}}$ $\epsilon_{v}$ $C_v$ [0.5ex] 0.1 2.51 0.026 0.019 0.0075 0.5 2.46 0.0883 0.06 0.026 0.9 2.65 0.45 0.25 0.080 1.9 2.3 0.67 0.49 0.1661 2.9 2.77 0.822 0.38 0.107 3.9 2.38 0.3137 0.27 0.10 4.5 2.59 0.263 0.1614 0.05
 $\sigma$ $\epsilon_{D}$ $\epsilon_{D^{'}}$ $\epsilon_{v}$ $C_v$ [0.5ex] 0.1 2.51 0.026 0.019 0.0075 0.5 2.46 0.0883 0.06 0.026 0.9 2.65 0.45 0.25 0.080 1.9 2.3 0.67 0.49 0.1661 2.9 2.77 0.822 0.38 0.107 3.9 2.38 0.3137 0.27 0.10 4.5 2.59 0.263 0.1614 0.05
 [1] Marc Homs-Dones. A generalization of the Babbage functional equation. Discrete & Continuous Dynamical Systems - A, 2021, 41 (2) : 899-919. doi: 10.3934/dcds.2020303 [2] Yi-Ming Tai, Zhengyang Zhang. Relaxation oscillations in a spruce-budworm interaction model with Holling's type II functional response. Discrete & Continuous Dynamical Systems - B, 2021  doi: 10.3934/dcdsb.2021027 [3] Maika Goto, Kazunori Kuwana, Yasuhide Uegata, Shigetoshi Yazaki. A method how to determine parameters arising in a smoldering evolution equation by image segmentation for experiment's movies. Discrete & Continuous Dynamical Systems - S, 2021, 14 (3) : 881-891. doi: 10.3934/dcdss.2020233 [4] Jiaquan Liu, Xiangqing Liu, Zhi-Qiang Wang. Sign-changing solutions for a parameter-dependent quasilinear equation. Discrete & Continuous Dynamical Systems - S, 2020  doi: 10.3934/dcdss.2020454 [5] Chongyang Liu, Meijia Han, Zhaohua Gong, Kok Lay Teo. Robust parameter estimation for constrained time-delay systems with inexact measurements. Journal of Industrial & Management Optimization, 2021, 17 (1) : 317-337. doi: 10.3934/jimo.2019113 [6] Dominique Chapelle, Philippe Moireau, Patrick Le Tallec. Robust filtering for joint state-parameter estimation in distributed mechanical systems. Discrete & Continuous Dynamical Systems - A, 2009, 23 (1&2) : 65-84. doi: 10.3934/dcds.2009.23.65 [7] Sumit Arora, Manil T. Mohan, Jaydev Dabas. Approximate controllability of a Sobolev type impulsive functional evolution system in Banach spaces. Mathematical Control & Related Fields, 2020  doi: 10.3934/mcrf.2020049 [8] Shuang Liu, Yuan Lou. A functional approach towards eigenvalue problems associated with incompressible flow. Discrete & Continuous Dynamical Systems - A, 2020, 40 (6) : 3715-3736. doi: 10.3934/dcds.2020028 [9] Marion Darbas, Jérémy Heleine, Stephanie Lohrengel. Numerical resolution by the quasi-reversibility method of a data completion problem for Maxwell's equations. Inverse Problems & Imaging, 2020, 14 (6) : 1107-1133. doi: 10.3934/ipi.2020056 [10] Hong Fu, Mingwu Liu, Bo Chen. Supplier's investment in manufacturer's quality improvement with equity holding. Journal of Industrial & Management Optimization, 2021, 17 (2) : 649-668. doi: 10.3934/jimo.2019127 [11] Skyler Simmons. Stability of Broucke's isosceles orbit. Discrete & Continuous Dynamical Systems - A, 2021  doi: 10.3934/dcds.2021015 [12] François Ledrappier. Three problems solved by Sébastien Gouëzel. Journal of Modern Dynamics, 2020, 16: 373-387. doi: 10.3934/jmd.2020015 [13] Ugo Bessi. Another point of view on Kusuoka's measure. Discrete & Continuous Dynamical Systems - A, 2020  doi: 10.3934/dcds.2020404 [14] Bo Chen, Youde Wang. Global weak solutions for Landau-Lifshitz flows and heat flows associated to micromagnetic energy functional. Communications on Pure & Applied Analysis, 2021, 20 (1) : 319-338. doi: 10.3934/cpaa.2020268 [15] Kihoon Seong. Low regularity a priori estimates for the fourth order cubic nonlinear Schrödinger equation. Communications on Pure & Applied Analysis, 2020, 19 (12) : 5437-5473. doi: 10.3934/cpaa.2020247 [16] José Luis López. A quantum approach to Keller-Segel dynamics via a dissipative nonlinear Schrödinger equation. Discrete & Continuous Dynamical Systems - A, 2020  doi: 10.3934/dcds.2020376 [17] Claudianor O. Alves, Rodrigo C. M. Nemer, Sergio H. Monari Soares. The use of the Morse theory to estimate the number of nontrivial solutions of a nonlinear Schrödinger equation with a magnetic field. Communications on Pure & Applied Analysis, 2021, 20 (1) : 449-465. doi: 10.3934/cpaa.2020276 [18] Alex H. Ardila, Mykael Cardoso. Blow-up solutions and strong instability of ground states for the inhomogeneous nonlinear Schrödinger equation. Communications on Pure & Applied Analysis, 2021, 20 (1) : 101-119. doi: 10.3934/cpaa.2020259 [19] Xiaorui Wang, Genqi Xu, Hao Chen. Uniform stabilization of 1-D Schrödinger equation with internal difference-type control. Discrete & Continuous Dynamical Systems - B, 2021  doi: 10.3934/dcdsb.2021022 [20] Van Duong Dinh. Random data theory for the cubic fourth-order nonlinear Schrödinger equation. Communications on Pure & Applied Analysis, 2021, 20 (2) : 651-680. doi: 10.3934/cpaa.2020284

Impact Factor: