- Research
- Open access
- Published:
Reconstructing the initial state for the nonlinear system and analyzing its convergence
Advances in Difference Equations volume 2014, Article number: 82 (2014)
Abstract
An algorithm to approximate the initial state of a nonlinear system is described, and its convergence is also analyzed in detail. The forward and backward observers are used alternately and repeatedly to solve the approximation problem, and their nudging term can be proved close to zero. Then the convergence problem based on the observers derived by using semi-discretization and full-discretization in space is considered.
1 Introduction
It is important to estimate the initial state of a linear partial difference system based on the observations over given time interval in science and engineering such as in oceanography, meteorology, medical imaging and so on, see for [1]. In oceanography such problem is called data assimilation for instance [2, 3]. The problem has been introduced in the quasi-geostrophic model in oceanography successfully [4] and arose in medical imaging by impedance-acoustic tomography [5, 6]. More recently, the time reversal method has been applied in the context of infinite-dimensional systems to estimate the initial data; see [7, 8].
The standard nudging method for solving the approximation problem usually adds a relaxation term to the equations of the system to construct the forward observation. Similarly the backward observation is constructed by adding a relaxation term with opposite sign. In this paper, performing the forward and backward observers repeatedly, our algorithm can be obtained.
Firstly, the paper estimates the initial state of the inverse problems of the nonlinear distributed parameter system according to its input and output function measured over some finite time interval. The main idea is to repeatedly apply the same segment of data back and forth in sequence by constructing two observers called the forward and backward observer, respectively. Two observers are constructed by adding a relaxation term which goes to 0 to the state equations under certain conditions and works in forward and backward time, respectively.
Secondly, the paper considers the convergence analysis of the iterative algorithm for the nonlinear system. The analysis is fully based on the numerical analysis derived by using the semi-discretization and the full-discretization successively, and the algorithm is still based on the observers method to simplify the problems.
Let X and Y be Hilbert space, called the state space and output space, respectively. Let be the generator of a strongly continuous group T of isometries on X. Assume the operator called the observation operator and let be the bounded operator. The above operators describe the time reversible nonlinear system and the system is described by the following equation:
where z and y are called the state and output function, respectively. Such systems are often used as models of vibrating systems, electromagnetic phenomena or in quantum mechanics.
Firstly, our aim is to reconstruct the initial data of the system when the output function y on the known time interval is given.
The paper is organized as follows: The preliminary knowledge is introduced in Section 2. The initial state is estimated just by one step iteration and the convergence is described briefly in Section 3. The correlative conclusions is considered after iterating n times in Section 4. The convergence accuracy is analyzed in detail for the iteration method for the nonlinear system in Section 5. The numerical result is showed in Section 6.
2 Description
Definition 2.1 System (1.1)-(1.2) is said to be exactly observable in some time Ï„ if there exists , such that
If system (1.1)-(1.2) is exactly observable in some time Ï„, it is exactly observable in any time. The inequality (2.1) is called the observation or observability inequality (see [9]). That guarantees the initial state is uniquely determined by the observed quantity on . To solve the infinite-dimensional system in the paper, we assume that the system is well-posed, i.e. the system is exactly observable.
Definition 2.2 There exists an operator that generates an exponentially stable semigroup on X and another operator where denotes the analog of the space , such that
Then the pair is said to be (forward) estimatable (see [1]).
Definition 2.3 There exists an operator that generates an exponentially stable semigroup on X and another operator where denotes the analog of the space , such that
Then the pair is said to be backward estimatable (see [1]).
Proposition 2.4 Assume that A is the skew-adjoint operator and T is the unitary group generated by A, then the following assertions are equivalent:
-
(i)
is exactly observable.
-
(ii)
is forward estimatable.
-
(iii)
is backward estimatable.
Proof The equivalence is contained in Proposition 3.7 in [1]. □
3 Properties of one step iteration
Assume is estimatable. We can construct an observer as follows: is the state of the forward observer and it satisfies the differential equation
where is an arbitrary initial guess of which can be proved independent of the guess in the following text.
Define the estimation error by , then
Thus where denotes the semigroup generated by at the time .
By using the method of separation of variables to (3.1), we can obtain
Now suppose is backward estimatable. We can also construct a backward observer as follows: is the state of the backward observer and it satisfies the differential equation
Define the estimation error by , then
Thus where denotes the semigroup generated by at the time . Since , we have .
Similarly, we can also get the solution of (3.3):
Proposition 3.1 and are given by (3.1) and (3.3), let and , and K, , C are the symmetric definite positive matrices. Then for any , if K, are large enough, we have
where means that any eigenvalue of the matrices tends to infinity.
Proof Since K, , C are symmetric definite positive matrices, when K, are large enough, and are definite. By utilizing the Green formula to (3.2), we can obtain
Thus
Similarly, we can also prove
 □
It can be seen that and are totally independent of the initial condition of the system.
Theorem 3.2 Assume is backward estimatable, then
If we set and , we have , and
Proof From , we have
If , we have
By Proposition 3.7 in [1], we have , then
Using a Neumann series, we can obtain
where denotes n times of . □
The process that computes by using the forward observer (3.1) and then computes by using the backward observer (3.3) is just one step iteration. For accuracy, the repeated multiple iterations should be further concerned as the above one step iteration.
4 Properties of multiple iterations
Consider the iterative algorithm on repeated estimation cycles. For , suppose and define and as the solutions of the following systems, respectively:
where is an arbitrary initial guess of which is independent of the guess and denotes the value at the n th iteration.
By , from (3.2), (3.4), it is easy to obtain
and
According to the repeated iterations, and , we can get
By (4.2) and the above equation, if , we have
and for , we have
According to Proposition 3.1, we know that
Similarly, for , we can get
and
It can be seen that and are totally independent of the initial condition of the system.
Theorem 4.1 Assume is backward estimatable, then
and for , we have
Proof From Theorem 3.2 and , we know that
Since , the conclusion can easily be obtained. □
Theorem 4.2 Assume is backward estimatable, and set , then
Proof It is similar to the proof of Theorem 3.2. □
The above iterative algorithm on the nonlinear system has been proved to be convergent if the feedback term K is large enough. and in the forward and backward observers are also totally determined by the output function of the system. Thus the initial state can be approximated by the algorithm, but the accuracy analysis is still a problem.
5 Numerical convergence
In this section, the convergence accuracy based on the observers is treated according to the semi-discretization and full-discretization method.
Let be the skew-adjoint operator, i.e., , then is the self-adjoint operator, i.e., . If A is the skew-adjoint operator, we often choose H and equivalent to , i.e., and .
The system (1.1)-(1.2) can be rewritten as
Throughout the section, let and .
For simplicity, let . Then the forward and backward observers (3.1) and (3.3) can be expressed, respectively, as
According to Theorem 3.2, we can obtain the expression of the initial state
The system (5.3)-(5.4) can be easily rewritten in the general form
where for the forward observer (5.3), we set , , and , and for the backward observer (5.4), we set , , and .
Define the subspace with the norm () in X. By the relations of the domain, we can get the embedding relations of the domain with the corresponding forms of the norm,
According to the embedding properties, we can obtain the following relations of the norm. There exist , for , such that
In order to prove the corresponding convergence conclusions, some preparatory lemma, which can simplify the proof procedure, has to be proved firstly.
Lemma 5.1 The initial value problem (5.6) is given, there exists , such that
where .
Proof By (5.6), we can obtain
By the triangle inequality and the boundedness of , , B, the first conclusion can be obtained.
By (5.6), we can obtain
Similarly, by the triangle inequality, the boundedness of B, C, the embedding properties, and the first inequality, the second conclusion can also be obtained. □
5.1 Semi-discretization
In this section, let h be the mesh size and be the optimal truncation parameter. We can construct the finite-dimensional subspace of where denotes the domain of the operator .
Define the orthogonal projection operator . Denote M as a constant independently of Ï„, and suppose that there exist , and , such that for ,
The generalized solution of the system (5.6) on the Galerkin significance is to find satisfying
for all and , where .
Start from the Galerkin method to approximate the variation formulation (5.8), i.e., the semi-discretization method is to find the unique solution satisfying the variation formulation
for all and , where is the given approximation of in X, and is the corresponding approximation of F in .
Assume that is the corresponding approximation of y in , and are the Galerkin approximations of and , respectively, and is the approximation of .
Proposition 5.1 There exist , and , such that for and , we have
Proof For all , subtracting (5.9) from (5.8), we can obtain
Noting that is established for , thus we have
Let , thus
Since and (5.10), the above equation with can be rewritten as
By the boundedness of B, C, we have
Since , the integration is
By (5.7), Lemma 5.1, and the embedding property, there exist , , and , such that for and , we have
Then the integration of the inequality (5.11) can be rewritten as
Thus, after the calculation of the integration, the result can be obtained. □
By the conclusion, the error approximations of the semigroup , , and the operator can be derived.
Proposition 5.2 There exist , and , such that for , and , we have
Proof By the triangle inequality, we have
For the first term, by (5.7), the embedding property and , the term can be estimated as
For the second term, using mathematical induction, we can prove that
When , by the definition of and , we have
When and , let and , respectively, we have
which is exactly (5.6).
Thus using Proposition 5.1, we can derive the existence of , , and , such that for and , we have
For the first term of (5.14), using the above conclusion and the uniform boundedness of , we get
Similarly, for the second term of (5.14), using the above conclusion, (5.7), and the uniform boundedness of and , we get
Substituting into (5.14), consequently
which shows that (5.13) holds when .
Now suppose that (5.13) holds for (). Then for n, we have
which is exactly (5.13). Thus we obtain the result. □
Next we estimate the error in semi-discretization.
Theorem 5.3 There exist , , and , such that for and , we have
Proof Using (5.5) and , we can get
Therefore, we have
where we have set
The first term, by and , can be estimated as
Similarly, the second term, by Proposition 5.2, can be estimated as
For the third term, from Proposition 5.2 we know that is uniformly bounded, thus we have
For the first term of (5.17), with (5.5), (5.7), and the embedding property we have
For the second term of (5.17), to estimate it we apply twice Proposition 5.1 for the time reversed backward observer and the forward observer, respectively.
Firstly, when , we have , , and ,
Then, when , , , we have
Applying Lemma 5.1 , we get .
And we can easily obtain
Thus the second term of (5.17) can be estimated as
Therefore, substituting (5.18) and (5.19) into (5.17), we can obtain
Above all, substituting (5.15), (5.16), and (5.20) into (5.14), we can obtain
which implies the conclusion holds. □
The choice of will lead to an explicit error estimate which is just dependent on h, and the proper choice of is important. If we choose , then according to Theorem 5.3, we can get
5.2 Full-discretization
Divide the time interval into N subintervals and let the time step ().
Denote (), then .
By using the implicit Euler scheme at time with the previous Galerkin approximation (5.9), assume
Then the full-discretization problem is to find the solution such that
for all and , where is the given approximation of and is the corresponding approximation of in X.
Assume that is the corresponding approximation of in Y, and are the approximations of and , respectively, and is the approximation of .
The convergence analysis is similar to that in the semi-discretization, thus we can prove two main ingredients of the error estimation as in the semi-discretization.
Proposition 5.4 There exist , , and , such that for and , we have
Proof Expand into the Taylor series at time and denote the residual term of the first order Taylor expansion by , then
Namely,
By the relation (5.24), for all and , we can get
Substituting (5.8) and (5.24) into (5.25) at time , then
Noting that is established for , thus
We can also easily get
Let , therefore for , we can obtain
and for , we can obtain
Let , , by the definition of , the above identity can be rewritten as
Substituting (5.26) and (5.27) into (5.28) with , then
By , we have
Similarly, by the definition of , we can easily obtain
Using the boundedness of B, C and from (5.7), (5.29), (5.30), and (5.31), we can see that there exist , , and such that for all and , we have
By the definition of in and the mean value theorem, we can obtain
From the fundamental property of the norm, Lemma 5.1, (5.33), and the embedding property, we can obtain
By the definition of in X, for , we can obtain
Thus
And since B, C are bounded, we have
Hence, from (5.35) and (5.36), we can obtain
And by simple iterations we get
Substituting (5.32), (5.34), and (5.37) into (5.38) with , then
Therefore we get the conclusion. □
Proposition 5.5 There exist , , and , such that for , , and , we have
Proof By the triangle inequality, we have
For the first term, using (5.7), the embedding property and , the term can be estimated as
For the second term, using mathematical induction, we can prove that
When , by definition of and , we have
When and , let and , respectively, it follows that
which is exactly (5.6).
Thus using Proposition 5.4, we can derive the existence of , , and , such that for and , we have
For the first term of (5.40), using the above conclusion and the uniform boundedness of , we get
For the second term of (5.40), similarly, using the above conclusion, (5.7), and the uniform boundedness of , we get
Substituting into (5.40), consequently
which shows that (5.39) holds when .
Now suppose that (5.39) holds for (), then for n, we have
which is exactly (5.39). Thus from (5.38) and (5.39), we obtain the result. □
Next, let us estimate the error in full-discretization.
Theorem 5.6 There exist , , and , such that for and , we have
Proof By (5.5) and , we get
Therefore, we have
where we have set
The first term, by and , can be estimated as
Similarly, the second term, by Proposition 5.5, can be estimated as
For the third term, from Proposition 5.5 we know that is uniformly bounded, thus we have
For the first term of (5.44), with (5.5), (5.7), and the embedding property, we have
For the second term of (5.44), to estimate it we apply Proposition 5.4 twice for the time reversed backward observer and the forward observer, respectively, which is similar to (5.19). Therefore,
Thus, substituting the above inequality and (5.45) into (5.44), we can obtain
Substituting (5.42), (5.43), and (5.46) into (5.41), we can obtain
which implies the conclusion holds. □
The choice of will lead to an explicit error estimate which is just dependent on h, and the proper choice of is important. If we choose , according to Theorem 5.6, we can get
6 Examples
In the section we apply algorithm to reconstruct the initial state for the nonlinear equation, and the algorithms were developed under Matlab. Let () and . Given the state space and the output space . The operators , and B are defined by (), and .
We consider the following initial and boundary value problem:
The output function is
The corresponding observer system is
where is an arbitrary initial guess of which is taken to be zero.
In order to show the efficiency of the iterative algorithm, we consider the particular case where , , , , , , and the initial data to be recover is . We use a Crank-Nicolson scheme and quasi-reversible method of regularization inverse inversion to simulate the observer systems (6.1) and (6.2) from one iteration to multiple iterations in time combined with a finite difference space discretization. Figure 1 shows the initial state, Figure 2 shows the final evolution of the output function. After one forward and backward iteration, we can obtain Figure 3 and Figure 4, obviously the result is not accurate enough, and after five iterations we obtain Figure 5. Figure 5 shows that the recursive algorithm reconstructs the initial state as far as possible. The algorithm take the simplest system and still needs to be improved.
7 Conclusion
The above iterative algorithm by using the forward and backward observers may estimate the initial state of the inverse problems of the nonlinear system under certain conditions. The numerical convergence accuracy analysis based on observers in the semi-discretization and the full-discretization can also be obtained. The convergence analysis of and towards has been showed for the nonlinear system if the truncation parameters and are optimal. The estimate error we have got provides the admissible upper bound under which the convergence can be guaranteed. The innovation in the paper is that this paper introduces the algorithm more systematically and comprehensively and demonstrates it in detail. We need to work on more applications of the algorithm and on the accuracy.
References
Ramdani K, Tucsnak M, Weiss G: Recovering the initial state of an infinite-dimensional system using observers. Automatica 2010, 46(10):1616–1625. 10.1016/j.automatica.2010.06.032
Auroux D, Blum J: A nudging-based data assimilation method: the Back and Forth Nudging (BFN) algorithm. Nonlinear Process. Geophys. 2008, 15: 305–319. 10.5194/npg-15-305-2008
Auroux D, Blum J: Back and forth nudging algorithm for data assimilation problems. C. R. Math. Acad. Sci. Paris 2005, 340: 873–878. 10.1016/j.crma.2005.05.006
Hoke J, Anthes A: The initialization of numerical models by a dynamic initialization technique. Mon. Weather Rev. 1976, 104: 1551–1556. 10.1175/1520-0493(1976)104<1551:TIONMB>2.0.CO;2
Liu K: Locally distributed control and damping for the conservative system. SIAM J. Control Optim. 1997, 35(5):1574–1590. 10.1137/S0363012995284928
Tucanak M, Weiss G Birkhäuser Advanced Texts. In Observation and Control for Operator Semigroups. Birkhäuser, Basel; 2009.
Byungik K: Multiple valued iterative dynamics models of nonlinear discrete-time control dynamical systems with disturbance. J. Korean Math. Soc. 2013, 50: 17–39. 10.4134/JKMS.2013.50.1.017
Curtain RF, Zwart H Texts in Applied Mathematics 21. In An Introduction to Infinite-Dimensional Linear Systems Theory. Springer, New York; 1995.
Ramdani K, Takahashi T, Tenenbaum G, Tucsnak M: A spectral approach for the exact observability of infinite-dimensional systems with skew-adjoint generator. J. Funct. Anal. 2005, 226: 193–229. 10.1016/j.jfa.2005.02.009
Acknowledgements
This work was supported by the National Natural Science Foundation of China. Grant No. 51205286.
Author information
Authors and Affiliations
Corresponding author
Additional information
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
All authors contributed equally to the writing of this paper. All authors read and approved the final manuscript.
Authors’ original submitted files for images
Below are the links to the authors’ original submitted files for images.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 2.0 International License (https://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
About this article
Cite this article
Xie, W., Chang, Y. Reconstructing the initial state for the nonlinear system and analyzing its convergence. Adv Differ Equ 2014, 82 (2014). https://doi.org/10.1186/1687-1847-2014-82
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/1687-1847-2014-82