Iterative algorithm for solving heavy root of nonlinear equation

By constructing the optimal fourth-order iteration algorithm and combining the self-accelerator operator and the memory iteration algorithm, the problem of slow re-root solving speed of nonlinear equations is solved, which significantly improves the convergence speed and accuracy, especially in multiple root solutions.

CN118964793AInactive Publication Date: 2024-11-15NANTONG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411013577.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-26
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

In the prior art, the iterative algorithm for solving nonlinear equations has a slow convergence rate and lacks efficient algorithms. In particular, self-acceleration technology has fewer applications in re-root solution.

Method used

A class of optimal fourth-order iterative algorithms for solving nonlinear equations are constructed, and self-acceleration operators are introduced, and five self-acceleration operators, including Newton's interpolation and Hermit's interpolation method are combined to construct an iterative algorithm with memory. The application scope of the algorithm is expanded and the convergence speed is improved.

Benefits of technology

By introducing self-accelerator operators and memory iteration algorithms, the convergence speed of re-root solutions for nonlinear equations is significantly improved, and higher iteration accuracy and faster convergence speed are achieved, especially in multiple root solutions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118964793A_ABST
    Figure CN118964793A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of solving nonlinear equation heavy roots in a class mode, in particular to a class of iterative algorithms for solving nonlinear equation heavy roots, and constructs a class of optimal four-order iterative algorithms for solving nonlinear equation heavy roots. A self-acceleration operator is introduced into the algorithm, and a corresponding iterative algorithm with memory is constructed; five self-acceleration operators based on a Newton interpolation method, a Hermite interpolation method and the like are respectively applied to the above iterative algorithm with memory, and the convergence speed of the algorithm is further improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of solving multiple roots of nonlinear equations, and in particular to an iterative algorithm for solving multiple roots of nonlinear equations. Background Art

[0002] Nonlinear equations are a type of equation that often appears in practical problems. Solving the roots of nonlinear equations has a wide range of applications in scientific and engineering calculations, image analysis of meteorological data, and the design of aircraft, automobiles, and ships. For a nonlinear equation f(x) = 0, if the number x * Satisfy f(x * )=0, then x * is the root of the equation. Further, if f(x * )=f′(x * )=...=f (m-1) (x * )=0,f (m) (x * )≠0, we call x * is the m-th root of the equation, where m is the multiplicity of the root. In particular, m = 1 is a single root. Due to the high singularity at the multiple roots, the research results on iterative algorithms for solving multiple roots are much less than those for solving single roots.

[0003] Newton iteration method is the most classic method for calculating the roots of nonlinear equations. Its iteration format is as follows:

[0004]

[0005] Newton's iteration method is suitable for finding a single root of an equation, and only converges linearly when finding multiple roots. In order to obtain a higher-order algorithm, E. The following improved Newton iteration method is proposed [1] :

[0006]

[0007] Here, m is the multiplicity of the multiple roots of the equation to be found.

[0008] In recent years, more and more scholars have made further modifications to the improved Newton iteration method and obtained a higher-order iterative algorithm. [2-6] . Work by XYWu [7] Inspired by V.Kanwar et al. [8] The following second-order convergent iterative formula is considered:

[0009]

[0010] The advantage of this algorithm is that it allows f′(x n )=0, thus expanding the application scope of the algorithm.

[0011] Self-acceleration technology has been widely used in iterative algorithms for solving single roots. Many studies have shown that self-acceleration operators can effectively improve the convergence order of the original algorithm. However, there are few studies on iterative algorithms for solving multiple roots. [9] First, we study the self-accelerating operators based on Newton interpolation for solving repeated roots and apply them to F. Zafar et al.

[10] The multi-step iterative algorithm given effectively improves the convergence speed of the algorithm. Summary of the invention

[0012] The purpose of the present invention is to solve the shortcomings existing in the prior art, and to propose a class of iterative algorithms for solving repeated roots of nonlinear equations. An optimal fourth-order iterative algorithm for solving repeated roots of nonlinear equations is constructed; a self-accelerating operator is introduced into the algorithm to construct a corresponding iterative algorithm with memory; five self-accelerating operators based on Newton interpolation method and Hermite interpolation method are respectively applied to the above iterative algorithm with memory to further improve the convergence speed of the algorithm.

[0013] In order to achieve the above object, the present invention adopts the following technical solutions:

[0014] A class of iterative algorithms for solving multiple roots of nonlinear equations. The specific steps are as follows:

[0015] Step 1: Based on Construct an optimal-order two-step iterative algorithm for solving multiple roots of nonlinear equations;

[0016] Where m is the multiplicity of the root x* of the nonlinear equation f(x)=0, x n is the iteration result of step n, y n is the pre-correction result of step n;

[0017] Adding parameters a and b to the denominator allows the first-order derivative of the function to be zero at the iteration point during the iteration process, thereby expanding the application scope of the iterative algorithm.

[0018] Step 2: In order to obtain the optimal fourth-order convergence speed, parameters a and b need to satisfy a certain positive proportional relationship. Set parameter a as the acceleration operator a n , construct an iterative algorithm with memory:

[0019]

[0020] The function G(x) satisfies The weight function of a is used; in the iterative process, the information of the iteration point is used to continuously update a n , making it closer to This improves the convergence speed.

[0021] Preferably, in step 2, five self-acceleration operators a based on Newton interpolation method and Hermite interpolation method are respectively n Applied to the iterative algorithm with memory, the convergence speed of the original algorithm was increased to and 5th order; among them, the five self-accelerating operators a n They are as follows:

[0022] (1) in where f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and node x n-1 The m+1 order difference quotient of (a total of m+2 nodes).

[0023] (2) in where f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and the pre-corrected value y at step n-1 n-1 m+1 order difference quotient of (a total of m+2 nodes);

[0024] (3) in

[0025]

[0026] where f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and node x n-1 m+1 order difference quotient of (a total of m+2 nodes); Represents the function f with respect to m+1 nodes x n and 2 nodes x n-1 m+2 order difference quotient (total m+3 nodes);

[0027] (4) in

[0028]

[0029] where f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and node x n-1 m+1 order difference quotient of (a total of m+2 nodes); Represents the function f with respect to m+1 nodes x n , node x n-1 and the pre-corrected value y at step n-1 n-1 m+2 order difference quotient (total m+3 nodes);

[0030] (5) in

[0031]

[0032] where f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and the pre-corrected value y at step n-1 n-1 m+1 order difference quotient of (a total of m+2 nodes); Represents the function f with respect to m+1 nodes x n and 2 pre-corrected values ​​y of the n-1th step n-1 The m+2 order difference quotient of (a total of m+3 nodes).

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] The present invention constructs an optimal fourth-order iterative algorithm for solving repeated roots of nonlinear equations; introduces a self-accelerating operator into the algorithm to construct a corresponding iterative algorithm with memory; and applies five self-accelerating operators based on Newton interpolation method and Hermite interpolation method to the above iterative algorithm with memory to further improve the convergence speed of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0036] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings, so that those skilled in the art can better understand the advantages and features of the present invention, thereby making a clearer definition of the protection scope of the present invention. The embodiments described in the present invention are only a part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present invention.

[0037] Example:

[0038] like Figure 1 As shown, the mathematical model and process of the algorithm of the present invention are as follows:

[0039] First consider the following iteration form:

[0040]

[0041] in, Parameters a and b satisfy Under certain conditions, the algorithm can achieve the optimal fourth-order convergence speed.

[0042] Modify the parameters a and b in formula (1) to a n ,b n , and the following iterative method is obtained:

[0043]

[0044] Five self-accelerating operators based on Newton interpolation and Hermite interpolation are respectively n Applied to the above formula (2):

[0045] (1)

[0046] in Here f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and node x n-1 m+1 order difference quotient of (a total of m+2 nodes);

[0047] (2)

[0048] in Here f (m) (x n ) means that the function f iterates at point x at step n nThe m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and the pre-corrected value y at step n-1 n-1 m+1 order difference quotient of (a total of m+2 nodes);

[0049] (3)

[0050] in,

[0051]

[0052] Here f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and node x n-1 m+1 order difference quotient of (a total of m+2 nodes); Represents the function f with respect to m+1 nodes x n and 2 nodes x n-1 m+2 order difference quotient (total m+3 nodes);

[0053] (4)

[0054] in,

[0055]

[0056] Here f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and node x n-1 m+1 order difference quotient of (a total of m+2 nodes); Represents the function f with respect to m+1 nodes x n , node x n-1 and the pre-corrected value y at step n-1 n-1 m+2 order difference quotient (total m+3 nodes);

[0057] (5)

[0058] in,

[0059]

[0060] Here f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and the pre-corrected value y at step n-1 n-1 m+1 order difference quotient of (a total of m+2 nodes); Represents the function f with respect to m+1 nodes x n and 2 pre-corrected values ​​y of the n-1th step n-1 The m+2 order difference quotient of (a total of m+3 nodes).

[0061] Applied to the above-mentioned iterative algorithm with memory (2), the convergence speed of the original algorithm is improved to and 5th order.

[0062] Specifically, the algorithm is implemented in Matlab R2017a, and the calculation results are retained to 10240 significant digits. Five steps of iteration are performed for each calculation of repeated roots. The weight function that meets the conditions is selected as The corresponding fourth-order iterative method is:

[0063]

[0064] The iterative algorithm with memory is:

[0065]

[0066] in

[0067] Example 1

[0068] f1(x)=(e x +x-10) 2 There is a double root α=2.070579905, that is, m=2.

[0069] Example 2

[0070] f2(x)=(x 3 -5.22x 2 +9.0825x-5.2675)ln(x-0.75) has a triple root α=1.75.

[0071] Example 3

[0072] f3(x)=(x 2 -e x -3x+2) 2 There is a double root α = 0.257530285.

[0073] Example 4

[0074] f4(x)=(cosx-x) 2 There is a double root α = 0.739085133215.

[0075] Table 1 Numerical experimental results of formula (8) and formula (9) with self-acceleration operators (3) and (4)

[0076]

[0077]

[0078] Table 2 Numerical experimental results of algorithm (9) with self-accelerating operators (5), (6) and (7)

[0079]

[0080] It can be seen from the above four examples that the convergence order (COC) of the iterative algorithm with memory is greater than that of the algorithm without memory. At the same time, when iterating to the fifth step, the error is smaller and the solution accuracy is higher.

[0081] In summary, the present invention constructs an optimal fourth-order iterative algorithm for solving multiple roots of nonlinear equations; introduces a self-accelerating operator into the algorithm to construct a corresponding iterative algorithm with memory; and applies five self-accelerating operators based on Newton interpolation and Hermite interpolation to the above iterative algorithm with memory to further improve the convergence speed of the algorithm.

[0082] The description and practice disclosed in the present invention are easy to think and understand for ordinary technicians in the technical field, and several improvements and modifications can be made without departing from the principles of the present invention. Therefore, modifications or improvements made without departing from the spirit of the present invention should also be regarded as the protection scope of the present invention.

[0083] The literature is as follows:

[0084] [1]E. Uber unendlich viele Algorithmen zur derGleichungen.Math.Ann.,1870,2:317–365.

[0085] [2]X.Zhou,X.Chen,Y.Song.Constructing higher-order methodsforobtaining the multiple roots of nonlinear equations.Comput.Math.Appl.,2011,235:4199-4206.

[0086] [3]M.S.Petkovi L.D.Petkovi Construction and efficiency ofmultipoint root-ratio methods for finding multiple zeros.J Comput.Appl.Math.,2019,351:54-65.

[0087] [4]X.Zhou,X.Chen,Y.Song.Families of third and fourth order methodsformultiple roots of nonlinear equations.Appl.Math.Comput.,2013,219:6030-6038.

[0089] [5]B.Liu,X.Zhou.A new family of fourth-order methods formultipleroots of nonlinear equations.Nonlinear Anal.Model.Control.,2013,18:143-152.

[0091] [6]B.Neta,C.Chun,M.Scott.On the development of iterative methodsformultiple roots.Appl.Math.Comput.,2013,224:358-361.

[0092] [7]X.Wu.A new continuation Newton-like method and itsdeformation.Appl.Math.Comput.,2000,112:75-78.

[0094] [8]V.Kanwar,S.Bhatia.M.Kansal,New optimal class of higher-ordermethods for multiple roots,permitting f’(x)=0.Appl.Math.Comput.,2013,222:564-574.

[0096] [9]X.Zhou,B.Liu.Iterative methods for multiple roots with memoryusingself-accelerating technique.J.Comput.Appl.Math.,2023,428:115181.

[0098]

[10] F.Zafar,A.Cordero,J.R.Torregrosa.Stability analysis of afamily ofoptimal fourth-order methods for multiple roots.Numer.Algor.,2019,81:947-981。

Claims

1. An iterative algorithm for solving multiple roots of nonlinear equations, characterized in that: The specific steps are as follows: Step 1: Based on Construct an optimal-order two-step iterative algorithm for solving multiple roots of nonlinear equations; Where m is the multiplicity of the root x* of the nonlinear equation f(x)=0, x n is the iteration result of step n, y n is the pre-correction result of step n; Adding parameters a and b to the denominator allows the first-order derivative of the function to be zero at the iteration point during the iteration process; Step 2: Set parameter a to the acceleration operator a n , construct an iterative algorithm with memory: Among them, the function G(x) satisfies G(0)=0,G′(0)=1, The weight function of In the iterative process, the information of the iteration point is used to continuously update a n , making it closer to Thereby improving the convergence speed; where f (m) (x*) represents the m-th order derivative of the function f at the root x*.

2. The iterative algorithm for solving multiple roots of nonlinear equations according to claim 1, characterized in that: In step 2, five self-accelerating operators a based on Newton interpolation and Hermite interpolation are respectively n Applied to the iterative algorithm with memory, the convergence speed of the original algorithm was increased to and 5th order; among them, the five self-accelerating operators a n They are as follows: (1) in where f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and node x n-1 m+1 order difference quotient of (a total of m+2 nodes); (2) in where f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and the pre-corrected value y at step n-1 n-1 m+1 order difference quotient of (a total of m+2 nodes); (3) in where f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and node x n-1 m+1 order difference quotient of (a total of m+2 nodes); Represents the function f with respect to m+1 nodes x n and 2 nodes x n-1 m+2 order difference quotient (total m+3 nodes); (4) in where f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and node x n-1 m+1 order difference quotient of (a total of m+2 nodes); Represents the function f with respect to m+1 nodes x n , node x n-1 and the pre-corrected value y at step n-1 n-1 m+2 order difference quotient (total m+3 nodes); (5) in where f (m) (x n ) means that the function f iterates at point x at step n n The m-th derivative at ; (m+1)! represents the factorial of (m+1); Represents the function f with respect to m+1 nodes x n and the pre-corrected value y at step n-1 n-1 m+1 order difference quotient of (a total of m+2 nodes); Represents the function f with respect to m+1 nodes x n and 2 pre-corrected values ​​y of the n-1th step n-1 The m+2 order difference quotient of (a total of m+3 nodes).