Previous Article in Journal
Impacts of Online and Offline Channel Structures on Two-Period Supply Chains with Strategic Consumers
Previous Article in Special Issue
Generalized High-Order Classes for Solving Nonlinear Systems and Their Applications

Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

# Extending the Applicability of Stirling’s Method

by
Cristina Amorós
1,
Ioannis K. Argyros
2,
Á. Alberto Magreñán
3,
Samundra Regmi
2,
Rubén González
1,* and
Juan Antonio Sicilia
1
1
Escuela Superior de Ingeniería y Tecnología, Universidad Internacional de La Rioja, 26006 Logroño, Spain
2
Department of Mathematics Sciences, Cameron University, Lawton, OK 73505, USA
3
Departamento de Matemáticas y Computación, Universidad de La Rioja, 26004 Logroño, Spain
*
Author to whom correspondence should be addressed.
Mathematics 2020, 8(1), 35; https://doi.org/10.3390/math8010035
Submission received: 30 November 2019 / Revised: 26 December 2019 / Accepted: 28 December 2019 / Published: 31 December 2019
(This article belongs to the Special Issue Multipoint Methods for the Solution of Nonlinear Equations)

## Abstract

:
Stirling’s method is considered as an alternative to Newton’s method when the latter fails to converge to a solution of a nonlinear equation. Both methods converge quadratically under similar convergence criteria and require the same computational effort. However, Stirling’s method has shortcomings too. In particular, contractive conditions are assumed to show convergence. However, these conditions limit its applicability. The novelty of our paper lies in the fact that our convergence criteria do not require contractive conditions. Hence, we extend its applicability of Stirling’s method. Numerical examples illustrate our new findings.

## 1. Introduction

In this work we deal with finding a fixed point $x *$ of the equation
$x = F ( x ) ,$
where F is a Fréchet-differentiable operator defined on a convex subset D of a Banach space X with values into itself. By I we denote the identity linear operator in $L ( X , X )$. The symbol $L ( X , X )$ stands for the space of bounded linear operators from X into X.
Many applications from different areas, including education, reduce to dealing with Equation (1) utilizing mathematical modelling [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24]. However, the solution $x *$ is found in closed form only in rare cases. This problem leads to the usage of methods that are iterative in nature.
We study Stirling’s method given for all $n = 0 , 1 , 2 , …$ by
$x n + 1 = x n − ( I − F ′ ( F ( x n ) ) ) − 1 ( x n − F ( x n ) ) ,$
where $x 0 ∈ D$. Further we will introduce an operator $Γ ( x ) ∈ L ( X , X )$ such that $Γ ( x ) = ( I − F ′ ( F ( x ) ) ) − 1$ with $x ∈ D$, and denote $Γ 0 = Γ ( x 0 )$ for use in later Sections.
This method converges quadratically as Newton’s method does, and also requires the same computational effort (see details in [1,22]). It is considered to be a useful alternative in cases where Newton’s method fails to converge (see such examples in [22]). However, the usage of Stirling’s method has a drawback, since the convergence criteria require contractions. We have detected some other problems listed in Remarks 3 and 4. These drawbacks limit the applicability of Stirling’s method. In order to extend its applicability, we do not use contractive conditions in our semi-local as well as the local convergence results.
The rest of the work is structured as follows. Section 2 includes the semi-local convergence analysis. Section 3 contains the local analysis. The numerical results are given in Section 4.

## 2. Semi-Local Convergence Analysis

Let $L 0 > 0 , L > 0$ and $γ ≥ 0$. Consider a real sequence ${ t n }$ as
$t 0 = 0 , t 1 = γ , t n + 2 = t n + 1 + L ( t n + 1 − t n ) 2 2 ( 1 − L 0 t n + 1 ) .$
Next, we study the convergence of sequence ${ t n }$ by developing relevant lemmas and theorems.
Lemma 1.
Suppose that
$h = L 1 γ < 1 2 ,$
where
$L 1 = 1 8 ( L + 4 L 0 + L 2 + 8 L 0 L ) .$
Then, sequence ${ t n }$ generated for $t 0 = 0$ by (4) is increasing, converges to its unique least upper bound $t *$, so that
$d 1 ≤ t * ≤ d 2 ,$
where
$d 1 = 1 − exp [ − L 0 γ ( 1 − L 0 γ ) ( 1 − δ L 0 γ ) ] L 0$
$d 2 = 1 − exp [ 2 L 0 γ 2 − L 0 γ + 2 δ 1 2 − δ 1 ] L 0 ,$
$δ = L 2 L 0 ( 1 − L 0 γ ) 2$
and
$δ 1 = L 2 L 0 ( δ 1 − δ ) 2 .$
Proof.
It is convenient to first simplify sequence ${ t n }$. Define sequence ${ α n }$ by $α n = 1 − L 0 t n$. Then, by (4) we can write $α 0 = 1 , α 1 = 1 − L 0 γ , α n + 1 = α n − L ( α n − α n − 1 ) 2 2 L 0 α n$. Moreover, define sequence ${ θ n }$ by $θ n = 1 − α n α n − 1$. Then, we can write $θ 1 = L 0 γ , θ n + 1 = L 2 L 0 ( θ n 1 − θ n ) 2$. We have by (4) that $δ θ 1 < 1$ and $0 < θ 2 < θ 1$. Suppose that $0 < θ k < θ k − 1$ and $δ θ k < 1$. Then, we get in turn that
$θ k + 1 = L 2 L 0 ( θ k 1 − θ k ) 2 < δ θ k 2 < θ k$
and
$δ θ k + 1 < δ θ k < 1 .$
Hence, ${ β n }$ is a decreasing sequence, so $α n = ( 1 − β n ) α n − 1$ and $t n = 1 − α n L 0$ are also decreasing sequences. In particular,
$α n = ( 1 − β n ) α n − 1 = … = ( 1 − β n ) … ( 1 − β 1 ) α 0 = ( 1 − β n ) … ( 1 − β 0 ) .$
From $0 < β 1 = L 0 γ < 1$, we get $0 < α n < 1$, so $t n = 1 − α n γ < 1 γ$. That is sequence ${ t n }$ is increasing, bounded from above by $1 L 0$, so it converges to $t *$.
Next, we show (4). We can write
$α * = lim n → ∞ α n = ∏ n = 1 ∞ ( 1 − β n ) ,$
or
$log 1 α * = ∑ n = 1 ∞ log 1 1 − β n .$
Using the estimate
$2 t − 1 t + 1 ≤ log t ≤ t 2 − 1 2 t for t > 1 ,$
we get first an upper bound for $log 1 α *$ by (5) and (6) and the inequality $2 n ≥ n + 1$ for $n = 0 , 1 , 2 …$:
$log 1 α * ≤ ∑ n = 1 ∞ β n ( 2 − β n ) 2 ( 1 − β n ) ≤ 1 1 − β 1 ∑ n = 0 ∞ β n + 1 ≤ 1 δ ( 1 − β 1 ) ∑ n = 1 ∞ ( δ θ 1 ) 2 n ≤ 1 δ ( 1 − β 1 ) ∑ n = 1 ∞ ( δ β 1 ) n = β 1 ( 1 − β 1 ) ( 1 − δ β 1 ) ,$
which together with $t * = 1 − α * L 0$ imply $t * ≤ d 2$. The lower bound in (4) is obtained similarly from the estimate:
$log 1 α * ≥ 2 ∑ n = 1 ∞ α n 2 − α n > 2 α 1 2 − α 1 + 2 α 2 2 − α 2 .$
Lemma 2.
Suppose that
$h = 1 2 .$
Then, sequence ${ t n }$ is increasingly converging to $1 L 0$.
Proof.
We have $α n = ( 1 − L 0 R ) n , β n = L 0 γ$ and $t n = 1 − ( 1 − L 0 γ ) n L 0$. Then, by (8), we get $0 ≤ L 0 γ < 1$.
In what follows the set denoted by $U ( x , a )$ is a ball with center $x ∈ X$ and of radius $a > 0$.
To simplify, the notation, by $| | | |$ in this work, we denote the operator norm or the norm on the Banach space. The semi-local convergence analysis is based on the conditions ($C$):
($C 1$)
$F : D ⊂ X → X$ is a Fréchet differentiable operator and there exist $x 0 ∈ D , c > 0 , γ ≥ 0$ such that $Γ 0 = ( I − F ′ ( F ( x 0 ) ) ) − 1 ∈ L ( X , X )$ with
$| | I − F ′ ( F ( x 0 ) ) | | ≤ c$
and
$| | Γ 0 ( x 0 − F ( x 0 ) ) | | ≤ γ .$
($C 2$)
There exist $a 0 ∈ [ 0 , 1 ) , b 0 > 0$ such that for each $x ∈ D$
$| | F ( x ) − F ( x 0 ) | | ≤ a 0 | | x − x 0 | |$
and
$| | Γ 0 ( F ′ ( F ( x ) ) − F ′ ( F ( x 0 ) ) | | ≤ b 0 | | F ( x ) − F ( x 0 ) | | .$
($C 3$)
Let $r 0 = 1 a 0 b 0$ and $D 0 = D ∩ U ( x 0 , r 0 )$. There exist $b > 0 , b 1 > 0$ such that for each $x , y ∈ D 0$
$| | Γ 0 ( F ′ ( x ) − F ( y ) ) | | ≤ b | | x − y | |$
and
$| | F ′ ( F ( x ) ) − F ′ ( F ( x 0 ) ) | | ≤ b 1 | | F ( x ) − F ( x 0 ) | | .$
($C 4$)
Hypotheses of Lemmas 1 and 2 hold with
$L = 2 b ( c + b 1 b 0 + 1 2 )$
and
$L 0 = a 0 b 0 .$
($C 5$)
$| | F ( x 0 ) − x 0 | | 1 − a 0 ≤ t * .$
($C 6$)
The ball $U ¯ ( x 0 , t * )$ is constructed such that
$U ¯ ( x 0 , t * ) ⊆ D .$
We suppose from now on that the conditions ($C$) hold.
Next, the semi-local convergence result is given for Stirling’s method (2).
Theorem 1.
Under conditions ($C$), sequence ${ x n }$ generated by Stirling’s method (2) is well defined, remains in $U ( x 0 , t * )$ for each $n = 0 , 1 , 2 , …$ and converges to $x * ∈ U ¯ ( x 0 , t * )$ which satisfies $x * = F ( x * )$ with Q-order of convergence 2. Moreover, the following estimates hold
$| | x n − x * | | ≤ t * − t n ,$
and $x *$ is the only fixed point of F in $U ( x 0 , t * * )$, with
$t * * = 2 b − ( 2 a 0 + 1 ) t * .$
Proof.
Let $x ∈ U ¯ ( x 0 , t * )$. We get by ($C 2$) and ($C 5$) that
$| | F ( x ) − x 0 | | ≤ | | F ( x ) − F ( x 0 ) | | + | | F ( x 0 ) − x 0 | | ≤ a 0 | | x − x 0 | | + | | F ( x 0 ) − x 0 | | ≤ a 0 t * + | | F ( x 0 ) − x 0 | | ≤ t * ,$
so $F ( x ) ∈ U ¯ ( x 0 , t * )$. Using ($C 2$) and the Lemmas 1 and 2, we have in turn that
$| | Γ 0 ( Γ ( x ) − Γ 0 ) | | = | | Γ 0 ( ( F ′ ( F ( x ) ) − F ′ ( F ( x 0 ) ) ) | | ≤ b 0 a 0 | | x − x 0 | | = L 0 | | x − x 0 | | ≤ L 0 t * < 1 .$
By the Lemma of Banach on invertible operators [21] (Perturbation Lemma 2.3.2, p. 45) $Γ ( x ) − 1 ∈ L ( X , X )$, and
$| | Γ ( x ) ( I − F ′ ( F ( x 0 ) ) ) | | ≤ 1 1 − L 0 | | x − x 0 | | .$
Using Stirling’s method (2):
$x k + 1 − F ( x k + 1 ) = F ( x k + 1 ) − F ( x k ) − F ( x k + 1 ) + F ( x k ) = F ′ ( y k ) ( x k + 1 − x k ) − ( F ( x k + 1 − F ( x k ) ) = ∫ 0 1 [ F ′ ( y k ) − F ′ ( x k + θ ( x k + 1 − x k ) ) ] ( x k + 1 − x k ) d θ .$
Then, in view of ($C 2$), ($C 3$) and Equation (11), we obtain in turn that
$| | Γ 0 ( x k + 1 − F ( x k + 1 ) ) | | ≤ b ∫ 0 1 | | y k − x k − θ ( x k + 1 − x k ) | | | | x k + 1 − x k | | d θ ≤ b [ | | y k − x k | | + 1 2 | | x k + 1 − x k | | ] | | x k + 1 − x k | | ≤ b [ ( | | I − F ′ ( F ( x 0 ) ) | | + | | F ′ ( y k ) − F ′ ( F ( x 0 ) ) | | ) + | | x k + 1 − x k | | + 1 2 | | x k + 1 − x k | | ] | | x k + 1 − x k | | ≤ b ( c + b 1 a 0 a 0 b 0 + 1 2 ) | | x k + 1 − x k | | 2 = L 2 | | x k + 1 − x k | | 2 .$
Next, we can connect the preceding estimates on sequence ${ x k }$ with ${ t k }$. Indeed, we get by ($C 1$) and Equation (3) that
$| | x 1 − x 0 | | = | | Γ 0 ( x 0 − F ( x 0 ) ) | | ≤ γ = t 1 = t 1 − t 0 .$
By induction, Equations (3), (4), (10) and (12), we have in turn that
$| | x k + 1 − x k | | = | | Γ k ( x k − F ( x k ) ) | | ≤ | | Γ k ( I − F ′ ( F ( x 0 ) ) ) | | | | Γ 0 ( x k − F ( x k ) ) | | ≤ L ( t k − t k − 1 ) 2 2 ( 1 − L 0 t k ) = t k + 1 − t k .$
Hence, ${ t k }$ defined by Equation (3) is a majorizing sequence for ${ x k }$. By Lemmas 1 and 2, sequence ${ t k }$ is complete as convergent to $t *$. It then follows by Equation (13) that sequence ${ x k }$ is also complete so it converges to some $x * ∈ U ¯ ( x 0 , t * )$. By the estimate (see (12))
$| | Γ 0 ( x k + 1 − F ( x k + 1 ) ) | | ≤ L 2 | | x k + 1 − x k | | 2 ≤ L 2 ( t k + 1 − t k ) 2 ,$
we deduce that $x * = F ( x * )$ by letting $k → ∞$. Estimate $| | x n − x * | | ≤ t * − t n$ follows from Equation (13) and for $λ = L 2 ( 1 − L 0 t * )$, we get that
$| | x k + 1 − x k | | ≤ L 2 ( 1 − L 0 t k ) | | x k − x k − 1 | | 2 ≤ λ | | x k − x k − 1 | | 2 ,$
which implies that the Q-order convergence of Stirling’s method (2) is two. Furthermore, to show the uniqueness part, let $y * ∈ U ( x 0 , t * * )$ with $F ( y * ) = y *$. Define the operator Q by $Q = − ∫ 0 1 Γ 0 F ′ ( x * + θ ( y * − x * ) ) d θ .$ In view of ($C 2$) and ($C 3$), we obtain in turn that
$| | I − ( Γ 0 − Q ) | | = | | ∫ 0 1 Γ 0 [ F ′ ( x * + θ ( y * − x * ) ) − F ′ ( F ( x 0 ) ) ] d θ | | ≤ b ∫ 0 1 | | x * + θ ( y * − x * ) − F ( x 0 ) | | d θ ≤ b [ | | F ( x * ) − F ( x 0 ) | | + 1 2 | | x * − x 0 | | + 1 2 | | y * − x 0 | | ] ≤ b [ ( a 0 + 1 2 ) t * + 1 2 t * * ] < 1 .$
Then, by (15) $( Γ 0 − Q ) − 1 ∈ L ( X , X )$. Finally, we obtain $y * = x *$ using the identity
$0 = Γ 0 ( y * − F ( y * ) − x * + F ( x * ) ) = ( Γ 0 − Q ) ( y * − x * ) .$
Remark 1.
(a)
The Stirling’s method usual conditions corresponding to ($C 2$) (first condition) are given by [22]:
($C 2$)’
$| | F ′ ( x ) | | ≤ a$ for each $x ∈ D$ and $a ∈ [ 0 , 1 )$.
That is, operator F must be a contraction on D. Moreover, the convergence of Stirling’s method was shown in [22] under ($C 2$), $D 0 = D$ and $a ∈ ( 0 , 1 3 ]$. However, in the present study no such assumption is made. Hence, the applicability of Stirling’s method (2) is extended. Notice also that we can have $a 0 ≤ a , b 0 ≤ b$ and c can be chosen as $b = c b 1$.
(b)
Estimate (4) is similar to the sufficient convergence Kantorovich-type criteria for the semi-local convergence of Newton’s method given by us in [4]. However, the constants $b 0 ¯$ and $b ¯$ are the center-Lipschitz and Lipschitz constants for operator F (see also part (e)).
(c)
If set $D 0$ is switched by $D 1 = D ∩ U ( x 1 , r 0 − | | x 0 − F ( x 0 ) | | )$, since $D 1 ⊆ D$ and the iterates remain in $D 1$ the results can be improved even further. The corresponding constants to b and $b 1$ will be at least as small.
(d)
In view of the proof of Theorem 1, scalar sequence ${ s n }$ defined by
$s 0 = 0 , s 1 = R , s n + 1 = s n + k n ( s n − s n − 1 ) 2 1 − L 0 s n ,$
is also a majorizing sequence for Stirling’s method (2), where
$k n = a b ( c + b 1 a 0 s n + 1 2 ) < L$
$s n ≤ t n ,$
$s n + 1 − s n ≤ t n + 1 − t n$
and
$s * = lim n → ∞ s n ≤ t * .$
(e)
Newton’s method for Equation (1) is given for all $n = 0 , 1 , 2 , …$ by
$y n + 1 = y n − ( I − F ′ ( y n ) ) − 1 ( y n − F ( y n ) ) .$
Consider, items $c ¯$, $γ ¯$, $L 0 ¯$, $L ¯$, $L 1 ¯$, $Γ 0 ¯$, $b 0 ¯$, $b ¯$, $b 1 ¯$, $r 0 ¯$, $D 0 ¯$ and $h ¯$, corresponding to c, γ, $L 0$, L, $L 1$, $Γ 0$, $b 0$, b, $b 1$, $r 0$, $D 0$ and h respectively as
$| | I − F ′ ( x 0 ) | | ≤ c ¯ ,$
$| | Γ 0 ¯ ( x 0 − F ( x 0 ) ) | | ≤ γ ¯ ,$
$| | Γ 0 ¯ ( F ′ ( x ) − F ′ ( x 0 ) ) | | ≤ b 0 ¯ | | x − x 0 | | ,$
$| | Γ 0 ¯ ( F ′ ( x ) − F ′ ( y ) ) | | ≤ b ¯ | | x − y | | ,$
$L ¯ = 2 b ¯ ( c ¯ + b 1 ¯ b 0 ¯ + 1 2 ) , b 1 ¯ = b 0 ¯ ,$
$L 0 ¯ = a 0 ¯ b 0 ¯ ,$
$r 0 ¯ = 1 a 0 ¯ b 0 ¯ ,$
$D 0 ¯ = D ∩ U ( x 0 , r 0 ¯ ) ,$
and
$h ¯ = L 1 ¯ , γ ¯ ≤ 1 2 ,$
where
$L 1 ¯ = 1 8 ( L ¯ + 4 L 0 ¯ + L ¯ 2 + 8 L 0 ¯ L ¯ ) .$
The scalar sequence $t ¯ n$ is defined as
$t 0 ¯ = 0 , t 1 ¯ = γ ¯ , t ¯ n + 2 = t ¯ n + 1 + L ¯ ( t ¯ n + 1 − t ¯ n ) 2 2 ( 1 − L 0 ¯ t ¯ n + 1 )$
Then, Stirling’s method sufficient convergence criteria, error bounds and information on the uniqueness of the solution are better than Newton’s method when the "bar" constants and sets are smaller than the non bar constants. Similar favorable comparison can be made in the local convergence case that follows.

## 3. Local Convergence

The conditions ($H$) are used in the local convergence analysis of Stirling’s method (2):
($H 1$)
$F : D ⊂ X → X$ is a Fréchet differentiable operator, and there exists $x * ∈ D$ such that $Γ * = ( I − F ′ ( x * ) ) − 1 ∈ L ( X , X )$ and $F ( x * ) = x *$.
($H 2$)
There exist $μ ∈ ( 0 , 1 ) , ξ 0 > 0$ such that for each $x ∈ D$
$| | F ( x ) − F ( x * ) | | ≤ μ | | x − x * | |$
and
$| | Γ * ( F ′ ( F ( x ) ) − F ′ ( F ( x * ) ) ) | | ≤ ξ 0 | | F ( x ) − F ( x * ) | | .$
($H 3$)
Let $D 0 * = D ∩ U ( x * , R 0 ) , R 0 = 1 ξ 0 μ$. There exists $ξ > 0$ such that for each $x , y ∈ D 0 *$
$| | Γ * ( F ′ ( x ) − F ′ ( y ) ) | | ≤ ξ | | x − y | | .$
($H 4$)
The ball $U ¯ ( x * , R )$ is constructed such that $U ¯ ( x * , R ) ⊆ D$, where
$R = 1 ( μ + 1 2 ) ξ + μ ξ 0 .$
Theorem 2.
Suppose that conditions $( H )$ hold. Then, sequence ${ x n }$ generated for $x 0 ∈ U ( x * , R ) − { x * }$ by Stirling’s method (2) is well defined in $U ( x * , R )$, remains in $U ( x * , R )$ for each $n = 0 , 1 , 2 , …$ and converges to $x * ∈ U ¯ ( x * , R )$. Moreover, the following inequality holds
$| | x n + 1 − x * | | ≤ ξ ( μ + 1 2 ) | | x n − x * | | 2 1 − μ ξ 0 | | x n − x * | | .$
Furthermore, if $R 1 = 2 ξ$, $x *$ is the only fixed point of F on $U ( x * , R 1 )$.
Proof.
We shall show using mathematical induction that sequence ${ x n }$ is well defined, remains in $U ( x * , R )$ and converges to $x *$ so that (16) is satisfied. We have by ($H 1$) and ($H 2$) for $x ∈ U ( x * , R )$ that
$| | F ( x ) − x * | | = | | F ( x ) − F ( x * ) | | ≤ μ | | x − x * | | ≤ R ,$
so $F ( x ) ∈ U ( x * , R )$. Then by ($H 2$)
$| | Γ * ( I − F ′ ( F ( x ) ) ) − Γ * | | = | | Γ * ( F ′ ( F ( x ) ) − F ′ ( F ( x * ) ) ) | | ≤ ξ 0 | | F ( x ) − F ( x * ) | | ≤ ξ 0 μ | | x − x * | | ≤ ξ 0 μ R < 1 .$
Hence, $Γ ( x ) ∈ L ( X , X )$ and
$| | Γ ( x ) ( I − F ( F ( x * ) ) ) | | ≤ 1 1 − μ 0 ξ 0 | | x − x * | | .$
In particular, (18) holds for $x = x 0$, which shows that $x 1$ is well defined by Stirling’s method for $n = 0$. We can write by ($H 1$) that
$x 1 − x * = x 0 − x * − ( I − F ′ ( F ( x 0 ) ) ) − 1 ( x 0 − F ( x 0 ) ) = ( I − F ′ ( F ( x 0 ) ) ) − 1 [ F ( x 0 ) − F ( x * ) − F ′ ( F ( x 0 ) ) ( x 0 − x * ) ] = ( I − F ′ ( F ( x 0 ) ) ) − 1 [ ∫ 0 1 ( F ′ ( x * + θ ( x 0 − x * ) ) − F ′ ( F ( x 0 ) ) ( x 0 − x * ) d θ ] .$
We get in turn by ($H 2$) and ($H 3$)
$| | Γ * ∫ 0 1 ( F ′ ( x * + θ ( x 0 − x * ) ) − F ′ ( F ( x 0 ) ) ) ( x 0 − x * ) d θ | | ≤ ξ ∫ 0 1 | | x * + θ ( x 0 − x * ) − F ( x 0 ) | | | | x 0 − x * | | d θ ≤ ξ [ | | F ( x * ) − F ( x 0 ) | | + θ | | x 0 − x * | | ] d θ ≤ ξ ( μ + 1 2 ) | | x 0 − x * | |$
Then, by (18)–(20), we get that also
$| | x 1 − x * | | ≤ | | ( I − F ′ ( F ( x 0 ) ) ) − 1 Γ * | |$
$≤ ξ ( μ + 1 2 ) | | x 0 − x * | | 2 1 − μ ξ 0 | | x 0 − x * | | ≤ | | x 0 − x * | | < R ,$
so (16) holds for $n = 0$ and $x 1 ∈ U ( x * , R )$. Switch $x 0$ by $x k$ in the preceding estimates, we arrive at (16). In view of the estimate $| | x k + 1 − x * | | < | | x k − x * | | < R$, we conclude that $lim k → ∞ x k = x *$ and $x k + 1 ∈ U ( x * , R )$. Let $x 0 = x *$ in (15) to show the uniqueness part. □
Remark 2.
The local results in the literature use ($C 2$)’ and $D 0 * = D$. But ($H 2$) is weaker than ($C 2$)’. Hence, we extend the applicability of Stirling’s method (2) in the local case too.

## 4. Numerical Example with Concluding Remarks

In the next example, we compare Stirling’s method with Newton’s method.
Example 1.
Let $D = X = R$. Consider function F on D as
$F ( x ) = − 1 3 x , x ≤ 3 1 3 x 2 − 7 3 x + 3 , 3 ≤ x ≤ 4 1 3 ( x − 7 ) , x ≥ 4 .$
Clearly, the quadratic polynomial joins smoothly with the linear parts.
(I)
Semilocal case (i). If we choose $x 0 = 3$, we see that $x 1 = y 1 = x * = 0$. Moreover, the semi-local convergence criteria of Theorem 1 are satisfied (with $γ = 0$, $a 0 = 1 3$ and $c = 4 3$).
(II)
Local convergence criteria of Theorem 2 (with $μ = 1 3$, since the derivative of the quadratic polynomial satisfies $1 3 | 2 x − 7 | ≤ 1 3$ for all $x ∈ [ 3 , 4 ]$).
(III)
In Table 1 and Table 2 we present some cases in which Stirling’s method stands better than Newton’s one.
In the current study, we have successfully demonstrated our claims on Stirling’s method by focusing on very classic problems, but in the future we will consider studying other complex problems such us solving symmetric ordinary differential equations with a more favorable theory.

## Author Contributions

All authors have equally contributed to this work. All authors have read and agreed to the published version of the manuscript.

## Funding

Research supported in part by Séneca 20928/PI/18 and by MINECO PGC2018-095896-B-C21.

## Conflicts of Interest

The authors have no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

## References

1. Argyros, I.K. Stirling’s method and fixed points of nonlinear operator equations in Banach spaces. Bull. Inst. Math. Acad. Sin. 1995, 23, 13–20. [Google Scholar]
2. IArgyros, I.K. A new iterative method of asymptotic order 1 + $2$ for the computation of fixed points. Int. J. Comput. Math. 2005, 82, 1413–1428. [Google Scholar] [CrossRef]
3. Argyros, I.K.; Chen, D. Results on the Chebyshev method in Banach spaces. Proyecciones 1993, 12, 119–128. [Google Scholar] [CrossRef]
4. Argyros, I.K.; Hilout, S. Weaker conditions for the convergence of Newton’s method. J. Complex. 2012, 28, 364–387. [Google Scholar] [CrossRef] [Green Version]
5. Argyros, I.K. A unifying local–semilocal convergence analysis and applications for two–point Newton–like methods in Banach space. J. Math. Anal. Appl. 2004, 298, 374–397. [Google Scholar] [CrossRef] [Green Version]
6. Argyros, I.K.; Magreñán, Á.A. Iterative Methods and Their Dynamics with Applications: A Contemporary Study; CRC Press: Boca Ratón, FL, USA, 2017. [Google Scholar]
7. Burrows, A.; Lockwood, M.; Borowczak, M.; Janak, E.; Barber, B. Integrated STEM: Focus on Informal Education and Community Collaboration through Engineering. Educ. Sci. 2018, 8, 4. [Google Scholar] [CrossRef] [Green Version]
8. Cianciaruso, F.; De Pascale, E. Estimates of majorizing sequences in the Newton-Kantorovich method: A further improvement. J. Math. Anal. Appl. 2006, 332, 329–335. [Google Scholar] [CrossRef] [Green Version]
9. Cordero, A.; Gutiérrez, J.M.; Magreñán, Á.A.; Torregrosa, J.R. Stability analysis of a parametric family of iterative methods for solving nonlinear models. Appl. Math. Comput. 2016, 285, 26–40. [Google Scholar] [CrossRef]
10. Chun, C.; Stanica, P.; Neta, B. Third order family of methods in Banach spaces. Comput. Math. Appl. 2011, 61, 1665–1675. [Google Scholar] [CrossRef] [Green Version]
11. Grout, I. Remote Laboratories as a Means to Widen Participation in STEM Education. Educ. Sci. 2017, 7, 85. [Google Scholar] [CrossRef] [Green Version]
12. Hernández, M.Á. The Newton Method for Operators with H’older Continuous First Derivate. J. Optim. Theory Appl. 2001, 109, 631–648. [Google Scholar] [CrossRef]
13. Kantorovich, L.V.; Akilov, G.P. Functional Analysis; Pergamon Press: Oxford, UK, 1982. [Google Scholar]
14. Laforce, M.; Noble, E.; Blackwell, C. Problem-Based Learning (PBL) and Student Interest in STEM Careers: The Roles of Motivation and Ability Beliefs. Educ. Sci. 2017, 7, 92. [Google Scholar] [CrossRef] [Green Version]
15. LeTendre, G.; McGinnis, E.; Mitra, D.; Montgomery, R.; Pendola, A. The American Journal of Education: Challenges and opportunities in translational science and the grey area of academic. Rev. Esp. Pedag. 2018, 76, 413–435. [Google Scholar] [CrossRef]
16. Magreñán, Á.A. Different anomalies in a Jarratt family of iterative root-finding methods. Appl. Math. Comput. 2014, 233, 29–38. [Google Scholar]
17. Magreñán, Á.A.; Argyros, I.K. A new tool to study real dynamics: The convergence plane. Appl. Math. Comput. 2014, 248, 215–224. [Google Scholar] [CrossRef] [Green Version]
18. Magreñán, Á.A. On the local convergence and the dynamics of Chebyshev-Halley methods with six and eight order of convergence. J. Comput. Appl. Math. 2016, 298, 236–251. [Google Scholar] [CrossRef]
19. Magreñán, Á.A.; Gutiérrez, J.M. Real dynamics for damped Newton’s method applied to cubic polynomials. J. Comput. Appl. Math. 2015, 275, 527–538. [Google Scholar] [CrossRef]
20. Nakakoji, Y.; Wilson, R. First-Year Mathematics and Its Application to Science: Evidence of Transfer of Learning to Physics and Engineering. Educ. Sci. 2018, 8, 8. [Google Scholar] [CrossRef] [Green Version]
21. Ortega, J.M.; Rheinboldt, W.C. Iterative Solution of Nonlinear Equations in Several Variables; Academic Press: New York, NY, USA, 1970. [Google Scholar]
22. Rall, L.B. Convergence of Stirling’s method in Banach spaces. Aequ. Math. 1975, 12, 12–20. [Google Scholar] [CrossRef]
23. Rheinboldt, W.C. An adaptive continuation process for solving systems of nonlinear equations. Pol. Acad. Sci. Banach. Cent. Publ. 1978, 3, 129–142. [Google Scholar] [CrossRef]
24. Traub, J.F. Iterative Methods for the Solution of Equations; Prentice-Hall Series in Automatic Computation: Englewood Cliffs, NJ, USA, 1964. [Google Scholar]
Table 1. Iteration of Newton’s and Stirling’s method with different starting points.
Table 1. Iteration of Newton’s and Stirling’s method with different starting points.
IterationNewton’s MethodStirling’s Method
0$3.4975$$3.4975$
1$− 646.501$$0.0618766$
2$− 1.13687 × 10 − 13$0
Table 2. Iteration of Newton’s and Stirling’s method with different starting points.
Table 2. Iteration of Newton’s and Stirling’s method with different starting points.
IterationNewton’s MethodStirling’s Method
0$3.5$$3.5$
1$0.0625$
20

## Share and Cite

MDPI and ACS Style

Amorós, C.; Argyros, I.K.; Magreñán, Á.A.; Regmi, S.; González, R.; Sicilia, J.A. Extending the Applicability of Stirling’s Method. Mathematics 2020, 8, 35. https://doi.org/10.3390/math8010035

AMA Style

Amorós C, Argyros IK, Magreñán ÁA, Regmi S, González R, Sicilia JA. Extending the Applicability of Stirling’s Method. Mathematics. 2020; 8(1):35. https://doi.org/10.3390/math8010035

Chicago/Turabian Style

Amorós, Cristina, Ioannis K. Argyros, Á. Alberto Magreñán, Samundra Regmi, Rubén González, and Juan Antonio Sicilia. 2020. "Extending the Applicability of Stirling’s Method" Mathematics 8, no. 1: 35. https://doi.org/10.3390/math8010035

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.