Abstract
In this paper, we estimate the asymptotic orders of probabilistic and average widths of the compact embedding operators from the Sobolev space into () with the Gaussian measure.
MSC:41A10, 41A46, 42A61, 46C99.
Similar content being viewed by others
1 Introduction and main results
Problems of n-widths in the approximation theory have by now been studied in depth. A great deal of classical problems have been solved, and interesting new developments have appeared. For example, the problems of probabilistic, average and stochastic widths, which can reflect the behavior of function on the whole class and give information about the measure of the elements in the class that can be approximated to this or that degree, are the problems of this kind. For the results related to the probabilistic, average and stochastic widths, the reader may be referred to Sul’din [1, 2], Traub et al. [3], Maiorov [4–7], Mathé [8–12], Sun [13, 14], and Ritter [15]. The new developments in this direction can be found in Fang’s papers [16–21]. Moreover, Carl and Pajor [22] proved an inequality with respect to the Gelfand numbers of an operator u from into a Hilbert space, from which one can immediately derive the inequality related to the Kolmogorov numbers by the known duality. In this article we continue the previous works and prove the estimates of probabilistic widths of the diagonal operators from onto .
First, we recall some useful concepts. Let W be a bounded subset of a normed linear space X with the norm , and be an N-dimensional subspace of X. The quantity
where
is called the deviation of W from . It shows how well the ‘worst’ elements of W can be approximated by . The number
where runs through all possible linear subspaces of X of dimension at most N, is called the Kolmogorov’s N-width of W in X. Assume that W contains a Borel field ℬ consisting of open subsets of W and equipped with a probabilistic measure μ defined on ℬ. That is, μ is a σ-additive nonnegative function on ℬ, and . Let be an arbitrary number. The corresponding probabilistic Kolmogorov’s -width of a set W with a measure μ in the space X is defined by
where runs through all possible subsets in ℬ with measure . The p-average Kolmogorov’s N-width is defined by
where in (2) runs over all linear subspaces of X of dimension at most N. Let be an m-dimensional normed space of vectors , with a norm
Consider in the standard Gaussian measure , which is defined as
where G is any Borel subset in . Obviously, .
Denote by the ball of radius ρ in . Let .
Let , be arbitrary and be a linear invertible operator from onto . We define the probabilistic -width of the operator acting in space equipped with the Gaussian measure v in -norm:
where , .
Maiorov in [6] proved the following result.
Theorem A [6]
For , , , then
In [22], Carl and Pajor proved the following result with respect to Gelfand numbers of an operator with values in a Hilbert space.
Theorem B [22]
Let T be an operator from into a Hilbert space H. Then
for , , where is a universal constant.
Detailed facts about the usual widths, such as the Kolmogorov’s N-widths and N th Gelfand numbers (or Gelfand N-widths) of T were given in the books [23–26].
Remark 1
-
(a)
Theorem A shows the asymptotic expression of the probabilistic widths of the identity embedding from into , .
-
(b)
Theorem B gives the upper estimate of Gelfand numbers of operators from into a Hilbert space, and some of its striking applications in the geometry of Banach spaces and Rademacher processes can be found in [22]. By the dual relation, it is easy to obtain the similar upper estimate of the Kolmogorov’s N-widths of operators from into , i.e.,
-
(c)
Motivated by Theorems A and B, in general cases, here we investigate the asymptotic estimate of probabilistic widths for diagonal operators from onto , .
Now we are in a position to formulate our main results.
Theorem 1 For , , then
Theorem 2 For , , then
2 Proof of main results
In order to prove Theorems 1 and 2, we also need some auxiliary assertions.
Lemma 1 Let , and let be a bounded linear invertible operator from onto . Then, for any vector ,
Proof First, assume that is a diagonal operator of , i.e., , for any . Without loss of generality, assume that the sequence of the absolute of eigenvalues , , is arranged non-increasingly, i.e., . It is known that
Since v is invariant with respect to orthogonal transformation of , it suffices to prove the lemma for the vector . Let be arbitrary. We have
Here we use the inequality
From (3) we obtain the assertion of the lemma by .
Next, assume that is a symmetric transformation of , then there is an orthogonal matrix U of order m such that the matrix is a diagonal matrix. Since the Gaussian measure is invariant for orthogonal transformation, the result holds for symmetric transformation .
Finally, assume that is a general invertible linear transformation from onto , then there are two matrices U and S such that , where U is an orthogonal matrix and S is a positive definite symmetric matrix. As the same reason above, the result holds for the transformation .
Thus Lemma 1 is proved. □
The following inequality will be used (see [27]). For any integers N and m with , there exists a subspace H of of dimension such that for any ,
where is an absolute constant.
Let be a set. We introduce in another norm for the operator :
Lemma 2 For any and an arbitrary operator from onto , there exists a subset of with measure such that
where a and are absolute constants.
Proof Let and consider the -net
for the in -norm. Using the inequality , we estimate the cardinality of S:
where a is some absolute constant.
Consider the polyhedron . Let be the set of extremal points of Q. The set consists of vectors with k coordinates equal to and the remaining coordinates zero. This implies that , and hence .
Let . In we consider the set , where
Let be any point, and be a point closest to z in -norm. Then for some . From Lemma 1,
Using the definition of Q, we have . From this and the definition of and G,
Therefore from (6),
Since , it follows from the inequalities (4) and (7) that
Using Lemma 1 and the inequality (5), we can estimate the measure of G:
Thus, Lemma 2 is proved. □
Proof of Theorem 1 Using the duality in and Lemma 2, we have
where is the orthogonal complement of H and . The proof of Theorem 1 is completed. □
Let us proceed to the proof of Theorem 2. For this, we first prove four lemmas. We introduce a definition. For arbitrary , the ε-cardinality of a subset K of is defined to be
where
is the deviation of K from the set in .
Let λ be a Lebesgue measure in ℝ, normalized by the condition , where . We consider Kolmogorov’s -width of the ball B with measure λ in the -norm:
where the is as above, the infima are over all possible subsets of measure and all subspaces with .
Lemma 3 Suppose that is an arbitrary subset with measure . Then, for any ,
Proof Let be any number, and let H be any subspace of ℝ with such that
Let . We consider the set . Let be the maximal subset of Q such that for all . Clearly, by maximality is a -net of Q for . The balls are disjoint and all contained in . Therefore, taking volumes we can obtain
Hence, we have
By and (9), we have
that is,
Now, we need to establish . Since , we have from (8)
From the inequality (11), the definition of and (8), it follows that
Consequently, letting , we get that , which together with (10) completes the proof of Lemma 3. □
From the relation (see [28])
the balls satisfy the inequalities
where Γ is the Euler Γ-function, and , depend only on p.
To estimate from below, we now need another auxiliary result.
Lemma 4 If is a diagonal operator from onto , then
where , , are non-zero eigenvalues of the operator rearranged as usual so that is non-increasing and each eigenvalue is repeated according to its multiplicity.
Proof It is known that
Accordingly,
Obviously,
from which the result of Lemma 4 follows immediately. □
Lemma 5 If and , then
Proof We first establish the inequality
Indeed, suppose that (13) does not hold. Then, for and some set of points , by
we have obtained a contradiction.
In the sequel, we may as well assume that is a diagonal operator from onto . Using the inequality (13), (12) and Lemma 4, we have
Next, assume that is a symmetric transformation of , then there is an orthogonal matrix U of order m such that the matrix is a diagonal matrix. Since the Lebesgue measure is invariant for orthogonal transformation, the result holds for symmetric transformation .
Finally, in the general case, is a general invertible linear transformation from onto , then there are two matrices U and S such that , where U is an orthogonal matrix and S is a positive definite symmetric matrix. As the same reason above, the result holds for the transformation .
Thus, we complete the proof of Lemma 5. □
Lemma 6 If , then
Proof From Lemma 3 and Lemma 5, we get
Let . Taking the logarithm of the inequality (14), we get the inequality
for some constants and and N with . Lemma 6 is proved. □
Proof of Theorem 2 According to Lemma 6, for any and any subspace with , there is a set with Lebesgue measure such that
for any element . On the unit sphere , we consider the subset .
Let be a Lebesgue measure on the sphere . We prove that
Indeed, assume that
We introduce in a polar system of coordinates , where and , and consider in B the cone
Then
We have obtained a contradiction.
Consider the set , . Using the inequality (16), we estimate the Gaussian measure of :
A direct computation shows that for ,
where is some absolute constant. It follows from this and (17) that for and for any ,
For any element , we have from (15)
Since ℒ is an arbitrary subspace with , it follows from (18) and (19) that
Theorem 2 is a direct consequence of this. □
References
Sul’din P: Wiener measure and its applications to approximation theory I. Izv. Vysš. Učebn. Zaved., Mat. 1959, 6: 145–158. (in Russian)
Sul’din P: Wiener measure and its applications to approximation theory II. Izv. Vysš. Učebn. Zaved., Mat. 1960, 5: 165–179. (in Russian)
Traub JF, Wasilkowski GW, Woźniakowski H: Information-Based Complexity. Academic Press, New York; 1988.
Maiorov VE: Discretization of the problem of diameters. Usp. Mat. Nauk 1975, 30: 179–180.
Maiorov VE: On linear widths of Sobolev classes and chains of extremal subspaces. Mat. Sb. 1980, 113(115):437–463. English transl. in Math. USSR Sb. 41 (1982)
Maiorov VE:Kolmogorov’s -widths of the spaces of the smooth functions. Russ. Acad. Sci. Sb. Math. 1994, 79: 265–279.
Maiorov VE: Linear widths of function spaces equipped with the Gaussian measure. J. Approx. Theory 1994, 77: 74–88. 10.1006/jath.1994.1035
Mathé P: S -Numbers in information-based complexity. J. Complex. 1990, 6: 41–66. 10.1016/0885-064X(90)90011-2
Mathé P: Random approximation of Sobolev embeddings. J. Complex. 1991, 7: 261–281. 10.1016/0885-064X(91)90036-W
Mathé P: A minimax principle for the optimal error of Monte Carlo methods. Constr. Approx. 1993, 9: 23–29. 10.1007/BF01229334
Mathé P: On optimal random nets. J. Complex. 1993, 9: 171–180. 10.1006/jcom.1993.1012
Mathé, P: Approximation theory of stochastic numerical methods. Habilitationsschrift, Fachbereich Mathematik, Freie Universität Berlin, Berlin (1994)
Sun Y, Wang C: μ -Average n -widths on the Wiener space. J. Complex. 1994, 10: 428–436. 10.1006/jcom.1994.1022
Sun Y, Wang C: Average error bounds of best approximation of continuous functions on the Wiener space. J. Complex. 1995, 11: 74–104. 10.1006/jcom.1995.1003
Ritter K Lecture Notes in Math. 1733. In Average-Case Analysis of Numerical Problems. Springer, New York; 2000.
Fang G, Qian L: Approximation characteristics for diagonal operators in different computational settings. J. Approx. Theory 2006, 140: 178–190. 10.1016/j.jat.2005.12.005
Fang G, Qian L: Optimization on class of operator equations in the probabilistic case setting. Sci. China Ser. A 2007, 50: 100–104. 10.1007/s11425-007-2041-9
Fang G, Ye P: Probabilistic and average linear widths of Sobolev space with Gaussian measure. J. Complex. 2003, 19: 73–84. 10.1016/S0885-064X(02)00005-5
Fang G, Ye P:Probabilistic and average linear widths of Sobolev space with Gaussian measure in norm. Constr. Approx. 2004, 20: 159–172.
Chen G, Fang G: Probabilistic and average widths of multivariate Sobolev spaces with mixed derivative equipped with the Gaussian measure. J. Complex. 2004, 20: 858–875. 10.1016/j.jco.2004.04.001
Chen G, Fang G: Linear widths of multivariate function spaces equipped with the Gaussian measure. J. Approx. Theory 2005, 132: 77–96. 10.1016/j.jat.2004.10.012
Carl B, Pajor A: Gelfand numbers of operators with values in a Hilbert space. Invent. Math. 1988, 94: 479–504. 10.1007/BF01394273
Pinkus A: n-Widths in Approximation Theory. Springer, New York; 1985.
Carl B, Stephani I: Entropy, Compactness and the Approximation of Operators. Cambridge University Press, Cambridge; 1990.
Pietsch A: Operator Ideals. North-Holland, Amsterdam; 1980.
Pietsch A: Eigenvalues and s-Numbers. Geest und Portig, Leipzig; 1987.
Gluskin ED: Norm of random matrices and widths of finite-dimensional sets. Mat. Sb. 1983, 120(162):180–189. English transl. in Math. USSR Sb. 48 (1984)
Pisier G: The Volume of Convex Bodies and Banach Space Geometry. Cambridge University Press, Cambridge; 1989.
Acknowledgements
The authors thank the editor and the referees for their valuable suggestions to improve the quality of this paper. The present investigations was supported by the Natural Science Foundation of Inner Mongolia Province of China under Grant 2011MS0103.
Author information
Authors and Affiliations
Corresponding author
Additional information
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
Both authors completed the paper together. They also read and approved the final manuscript.
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
Zhou, J., Li, Y. Estimates of probabilistic widths of the diagonal operator of finite-dimensional sets with the Gaussian measure. J Inequal Appl 2013, 277 (2013). https://doi.org/10.1186/1029-242X-2013-277
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/1029-242X-2013-277