Abstract
In this paper, we construct an optimal quadrature formula in the sense of Sard by Sobolev’s method in the space. We give explicit expressions for the corresponding optimal coefficients. This formula is exact for exponentional–trigonometric functions.
MSC:
41A05; 41A15; 65D30; 65D32
1. Introduction
It is known that in many practical situations, especially in engineering, economics, and medicine, if you want to solve some problems, you will usually need to take the integral of a function. But we have difficulty in calculating the exact values of some definite integrals. So we need to construct optimal quadrature formulas to estimate their numerical values (see [1,2,3,4,5,6]). Additionally, quadrature formulas also play an important role in calculating the numerical solution to differential and integral equations. And the accuracy of the quadrature formulas is an important problem, too. Specific applications include improving the accuracy and stability of numerical solutions when solving partial differential equations numerically. In signal processing and image processing, optimal quadrature formulas can better handle the local and global properties of signals and images, enabling functions such as noise reduction and feature extraction. In the Monte Carlo method, its theory can also be combined to optimize sampling strategies and integral estimates, improving computational efficiency and accuracy for calculating high-dimensional integrals, probability distributions, and other problems. We need to continuously improve the algorithms to make the errors smaller and smaller. The specific algorithms in this paper can also be applied. For example, to construct an optimal quadrature formula, it is necessary to determine the optimal coefficients by minimizing the norm of the error functional. In numerical integration algorithms, the error estimation theory in Sobolev spaces can be used to improve the algorithm and enhance its accuracy. In function interpolation algorithms, finite element algorithms, and others, the idea of function approximation is also required. According to the function properties in Sobolev spaces, appropriate approximation functions and algorithm strategies are selected to achieve efficient calculation and processing of complex functions. Nikol’skii’s problem is the minimization of the norm of the error functional by coefficients and by nodes such as that in [3]. The minimization of the norm of the error functional with fixed nodes by coefficients is Sard’s problem. The corresponding solutions are called the optimal quadrature formulas in the sense of Nikol’skii and in the sense of Sard, respectively. Reducing errors and studying optimal quadrature formulas can help calculate the loss functions involved in the training and evaluation of machine learning models. This, in turn, can improve the generalization ability and prediction accuracy of the models, and has potential application value in fields such as image recognition and speech recognition.
Work [1] studies Sard’s problem of constructing optimal quadrature formulas in the space. So, in this paper, we will only consider the case where .
We denote a class of functions by defined on the interval , which possesses an absolutely continuous sixth derivative on and whose seventh derivative is in , under the pseudo-inner product
The class is a Hilbert space if we identify functions that differ by a solution of the equation . The norm
corresponds to the inner product (1).
In Sard’s problem, for a function from the space , we consider the following quadrature formula:
where are coefficients and are nodes situated in the interval , are given values, and N is a natural number. The error of the quadrature formula is
and is the value of the error functional l at the given function . The error functional l has the form
where is the characteristic function of the interval [0, 1], and is Dirac’s delta-function.
By the Cauchy–Schwarz inequality, the absolute value of the error (3) is estimated by the norm
of the error functional l as follows:
where is the conjugate space of the space .
The construction of optimal quadrature formulas in the space with fixed nodes and given is to find such coefficients to satisfy the equation
In short, we will find the minimum of the norm (5) of the error functional l by coefficients for fixed .
Thus, if we want to construct the optimal quadrature formulas in the space, we have to solve two problems. First, we need to calculate the norm of the error functional l for the given quadrature Formula (2). Second, find such values of the coefficients , which satisfy equality (6).
The rest of the paper is organized as follows: In Section 2, we use the extremal function of the error functional (4) to find the norm of the error functional and give a representation of the norm. In Section 3, we obtain the system of linear equations for coefficients of the optimal quadrature formulas in the space , and the existence and uniqueness of the solution of this system are to be proved. Section 4 includes the formulas of the coefficients of the optimal quadrature formulas. In this section, we obtain the optimal quadrature formulas in the sense of Sard in the space . In Section 5, we calculate the explicit coefficients of this problem.
2. The Extremal Function and Representation of the Norm of the Error Functional
In order to solve the first problem, we should use the extremal function of the given functional. So, the extremal function of the functional l is the function satisfying the equation
Since is a Hilbert space, using the Riesz theorem on the general form of a linear continuous functional on a Hilbert space, a unique extremal function can be found for the error functional . Then, for the functional and for any , there exists such a function satisfying equality
and . Here, is the inner product defined by equality (1) in the space .
Then, we will find the solution to the equation of equality (7). We consider a situation when . We have
Integrating by parts the right hand side of (7), we obtain
Taking into account the function is unique and Equation (8), we can obtain the following form:
with the boundary conditions
The solution to Equation (9) with the boundary conditions is as follows:
where
and
where , and are constants. We should impose the following conditions:
That means the quadrature formula (2) is exact for linear combinations of functions
for .
Then, using Equations (13)–(15), we obtain
where Y(x) is the function defined by (12). Using (10), we obtain
After some calculation, we obtain
Thus, the first problem is solved.
3. The System of Linear Equations for Coefficients
We need to find the minimization of the square of the norm (16) for the error functional l by coefficients . In order to find a conditional minimum point of function (16), we apply the Lagrange method. Then, we define and . We have the following function:
where are Lagrange multipliers.
Let the partial derivatives of the function by coefficients and by , and be equal to zero; we can obtain the following difference system of linear equations with unknowns:
where
and is defined by (11).
The solution to this system exists and is unique in the sense of Sard in a Hilbert space. Then, we consider the case of equally spaced nodes. Suppose
We also suppose when and . Then, we use the convolution of two discrete argument functions and and have the following form:
4. The Formulas of the Coefficients
In order to solve the system (21)–(24), we need to use the differential operator . We use the function of discrete argument which was constructed in [1] and satisfies the equation
where
and is the discrete delta-function and has the form
Following [1], we have
Theorem 1.
The discrete analog of the differential operator satisfying Equation (25) has the form
where and
are roots of the polynomial with , and h is a small positive parameter.
Theorem 2.
The discrete analog for satisfies the following equations:
Next, we introduce the following functions:
Then, for the coefficients of the optimal quadrature formula, we obtain
5. The Calculation About Optimal Coefficients
If we want to obtain the explicit value of the coefficients of the optimal quadrature formula, calculating in the manner of work [2], we have to calculate the convolution .
Theorem 3.
The convolution has the form
where
Proof.
We consider the convolution of two discrete functions and take Equation (27) into account. Then, we have
where
When , we have
For , we have
When , we have
Theorem 3 is proved. □
It is easy to see that if we want to solve this problem, we will consider the convolution . And in (28), there are 14 unknowns, and we also need to find their solutions. Next, we will find the expression for . In the present work, we obtain
where is defined by (26). After some calculations as work [4], we can obtain , which has following form:
Next, taking Theorem 2 into account, we consider the convolution . We have
Then, we can obtain the value of the convolution as the following form:
where
So, if we consider Theorem 3, we can obtain a simple form of the expression of in the sense of Sard in the space . The simple form is as follows:
In this form, we have 14 unknowns. To find out about them, we will take Equations (21)–(24). In the rest of the paper, we will first calculate the convolution . Then, use Equations (22) and (23) and the equality of the coefficients on the right and left sides of Equation (21) to construct a system of linear equations which includes 21 equations and 21 unknowns.
Next, we calculate the convolution .
where
After some calculation, we have
where
Then, taking Equation (21) into account, we obtain
where , and are unknowns. We can construct a system of linear equations which include 21 equations and 21 unknowns by this equation and (22)–(24). The system of linear equations has the following form:
where
Then, the second problem is solved. Finally, we obtain the optimal quadrature formula as the form (2) in the sense of Sard in the space .
Author Contributions
Resources, X.L.; writing—original draft, Y.Y.; writing—review and editing, X.L. All authors have read and agreed to the published version of the manuscript.
Funding
Initial Research Fund of North China University of Technology: 110051360002.
Data Availability Statement
Data are contained within the article.
Conflicts of Interest
The authors declare no conflicts of interest.
References
- Boltaev, A.; Hayotov, A.; Shadimetov, K. Construction of Optimal Quadrature Formulas Exact for Exponentional-trigonometric Functions by Sobolev’s Method. Acta Math. Sin. 2021, 37, 1066–1088. [Google Scholar] [CrossRef]
- Boltaev, A.K.; Hayotov, A.R.; Shadimetov, K.M. About coefficients and order of convergence of the optimal quadrature formula. Am. J. Numer. Anal. 2014, 2, 35–48. [Google Scholar]
- Fang, G.; Li, X. Optimal quadrature problem on classes defined by kernels satisfying certain oscillation properties. Numer. Math. 2006, 105, 133–158. [Google Scholar] [CrossRef]
- Shadimetov, K.M.; Hayotov, A.R. Computation of coefficients of optimal quadrature formulas in the space W2(m,m − 1)(0, 1). Uzbek. Math. Zh. 2004, 3, 67–82. (In Russian) [Google Scholar]
- Shadimetov, K.M.; Hayotov, A.R.; Nuraliev, F.A. Construction of optimal interpolation formulas in the Sobolev space. J. Math. Sci. 2002, 264, 782–793. [Google Scholar] [CrossRef]
- Shadimetov, K.M.; Mamatova, N. Optimal quadrature formulas with derivatives in a periodic space. AIP Conf. Proc. 2021, 2356, 020030. [Google Scholar]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).