# A superlinearly convergent hybrid algorithm for systems of nonlinear equations

Open AccessResearch

DOI: 10.1186/1029-242X-2012-180

Zheng, L. J Inequal Appl (2012) 2012: 180. doi:10.1186/1029-242X-2012-180

## Abstract

We propose a new algorithm for solving systems of nonlinear equations with convex constraints which combines elements of Newton, the proximal point, and the projection method. The convergence of the whole sequence is established under weaker conditions than the ones used in existing projection-type methods. We study the superlinear convergence rate of the new method if in addition a certain error bound condition holds. Preliminary numerical experiments show that our method is efficient.

MSC: 90C25, 90C30.

### Keywords

nonlinear equations projection method global convergence superlinear convergence

## 1 Introduction

Let be a continuous mapping and be a nonempty, closed, and convex set. The inner product and norm are denoted by and , respectively. Consider the problem of finding
(1.1)
Let S denote the solution set of (1.1). Throughout this paper, we assume that S is nonempty and F has the property that
(1.2)

The property (1.2) holds if F is monotone or more generally pseudomonotone on C in the sense of Karamardian [1].

Nonlinear equations have wide applications in reality. For example, many problems arising from chemical technology, economy, and communications can be transformed into nonlinear equations; see [2, 3, 4, 5]. In recent years, many numerical methods for problem (1.1) with smooth mapping F have been proposed. These methods include the Newton method, quasi-Newton method, Levenberg-Marquardt method, trust region method, and their variants; see [6, 7, 8, 9, 10, 11, 12, 13, 14].

Recently, the literature [15] proposed a hybrid method for solving problem (1.1), which combines the Newton, proximal point, and projection methodologies. The method possesses a very nice globally convergent property if F is monotone and continuous. Under the assumptions of differentiability and nonsingularity, locally superlinear convergence of the method is proved. However, the condition of nonsingularity is too strong. Relaxing the nonsingularity assumption, the literature [16] proposed a modified version for the method by changing the projection way, and showed that under the local error bound condition which is weaker than nonsingularity, the proposed method converges superlinearly to the solution of problem (1.1). The numerical performances given in [16] show that the method is really efficient. However, the literatures [15, 16] need the mapping F to be monotone, which seems too stringent a requirement for the purpose of ensuring global convergence property and locally superlinear convergence of the hybrid method.

To further relax the assumption of monotonicity of F, in this paper, we propose a new hybrid algorithm for problem (1.1) which covers one in [16]. The global convergence of our method needs only to assume that F satisfies the property (1.2), which is much weaker than monotone or more generally pseudomonotone. We also discuss the superlinear convergence of our method under mild conditions. Preliminary numerical experiments show that our method is efficient.

## 2 Preliminaries and algorithms

For a nonempty, closed, and convex set and a vector , the projection of x onto Ω is defined as

We have the following property on the projection operator; see [17].

Lemma 2.1Letbe a closed convex set. Then it holds that

Algorithm 2.1 Choose , parameters , λ, , , , , , and set .

Step 1. Compute . If , stop. Otherwise, let , . Choose a positive semidefinite matrix . Compute such that
(2.1)
where
(2.2)

Stop if . Otherwise,

Step 2. Compute , where and is the smallest nonnegative integer m satisfying
(2.3)
Step 3. Compute
where and

Remark 2.1 When we take parameters , , and the search direction , our algorithm degrades into one in [16]. At this step of getting the next iterate, our projection way and projection region are also different from the one in [15].

Now we analyze the feasibility of Algorithm 2.1. It is obvious that satisfying conditions (2.1) and (2.2) exists. In fact, when we take , satisfies (2.1) and (2.2). Next, we need only to show the feasibility of (2.3).

Lemma 2.2For all nonnegative integerk, there exists a nonnegative integersatisfying (2.3).

Proof If , then it follows from (2.1) and (2.2) that , which means Algorithm 2.1 terminates with being a solution of problem (1.1).

Now, we assume that , for all k. By the definition of , the Cauchy-Schwarz inequality and the positive semidefiniteness of , we have
(2.5)
Suppose that the conclusion of Lemma 2.2 does not hold. Then there exists a nonnegative integer such that (2.3) is not satisfied for any nonnegative integer m, i.e.,
(2.6)
Letting and by the continuity of F, we have

Which, together with (2.5), , and , we conclude that , which contradicts . This completes the proof. □

## 3 Convergence analysis

In this section, we first prove two lemmas, and then analyze the global convergence of Algorithm 2.1.

Lemma 3.1If the sequencesandare generated by Algorithm  2.1, is bounded andFis continuous, thenis also bounded.

Proof Combining inequality (2.5) with the Cauchy-Schwarz inequality, we obtain
By the definition of and , it follows that

From the boundedness of and the continuity of F, we conclude that is bounded, and hence so is . This completes the proof. □

Lemma 3.2Letbe a solution of problem (1.1) and the functionbe defined by (2.4). If condition (1.2) holds, then
(3.1)

In particular, if, then.

Proof
(3.2)
(3.3)
where the inequality follows from (2.3).

where the inequality follows from condition (1.2) and the definition of .

If , then because . The proof is completed. □

Remark 3.1 Lemma 3.2 means that the hyperplane

strictly separates the current iterate from the solution set of problem (1.1).

Let and . Since

where the first inequality follows from condition (1.2), the second one follows from (2.3), and the last one follows , which shows that is a descent direction of the function at the point .

We next prove our main result. Certainly, if Algorithm 2.1 terminates at Step k, then is a solution of problem (1.1). Therefore, in the following analysis, we assume that Algorithm 2.1 always generates an infinite sequence.

Theorem 3.1IfFis continuous onC, condition (1.2) holds and, then the sequencegenerated by Algorithm  2.1 globally converges to a solution of (1.1).

Proof Let be a solution of problem (1.1). Since , it follows from Lemma 2.1 that
i.e.,
which shows that the sequence is nonincreasing, and hence is a convergent sequence. Therefore, is bounded and
(3.4)
From Lemma 3.1 and the continuity of F, we have that is bounded; that is, there exists a positive constant M such that
By (2.3) and the choices of and λ, we have
This, together with inequality (3.4), we deduce that

Now, we consider the following two possible cases:

Suppose first that . Then we must have
From the definition of , the choice of and , each case of them follows that
Since F is continuous and is bounded, which implies that the sequence has some accumulation point such that

This shows that is a solution of problem (1.1). Replacing by in the preceding argument, we obtain that the sequence is nonincreasing, and hence converges. Since is an accumulation point of , some subsequence of converges to zero, which implies that the whole sequence converges to zero, and hence .

Suppose now that . Let be any accumulation point of and be the corresponding subsequence converging to . By the choice of , (2.3) implies that
Since F is continuous, we obtain by letting that
(3.5)

From (2.5) and (3.5), we conclude that , which contradicts . Hence, the case of is not possible. This completes the proof. □

Remark 3.2 Compared to the conditions of the global convergence used in literatures [15, 16], our conditions are weaker.

## 4 Convergence rate

In this section, we provide a result on the convergence rate of the iterative sequence generated by Algorithm 2.1. To establish this result, we need the following conditions (4.1) and (4.2).

For , there are positive constants δ, , and such that
(4.1)
and
(4.2)
where denotes the distance from x to solution set S, and
If F is differentiable and is locally Lipschitz continuous with modulus , then there exists a constant such that
(4.3)
In fact, by the mean value theorem of vector valued function, we have
where . Under assumptions (4.2) or (4.3), it is readily shown that there exists a constant such that
(4.4)

In 1998, the literature [15] showed that their proposed method converged superlinearly when the underlying function F is monotone, differentiable with being nonsingular, and ∇F is locally Lipschitz continuous. It is known that the local error bound condition given in (4.1) is weaker than the nonsingular. Recently, under conditions (4.1), (4.2), and the underlying function F being monotone and continuous, the literature [16] obtained the locally superlinear rate of convergence of the proposed method.

Next, we analyze the superlinear convergence rate of the iterative sequence under a weaker condition. In the rest of section, we assume that , , where .

Lemma 4.1Letbe a positive semidefinite matrix and. Then

(1) ;

(2) .

Proof See [18]. □

Lemma 4.2Suppose thatFis continuous and satisfies conditions (1.2), (4.1), and (4.2). If there exists a positive constantNsuch thatfor allk, then for allksufficiently large,

(1) ;

(2) , where, andare all positive constants.

Proof For (1), let and be the closest solution to . We have
i.e., . Thus, by (2.1), (2.2), (4.2), and Lemma 4.1, we have
By and , it follows that
From (4.1) and the choice of , it holds that
From the boundedness of , there exists a positive constant such that
Therefore,
(4.5)

We obtain that the left-hand side of (1) by setting .

For the right-hand side part, it follows from (2.1) and (2.2) that

We obtain the right-hand side part by setting .

For (2), using (2.1) and (2.2), we have

By setting , we obtain the desired result. □

Lemma 4.3Suppose that the assumptions in Lemma  4.2 hold. Then for allksufficiently large, it holds that
Proof By and the continuity of F, we have
By Lemma 4.2(1), we obtain that
which means that for all k sufficiently large. Hence, it follows from (4.2) that
(4.6)
where . Using (2.1) and (2.2), (4.6) can be written as
(4.7)
Hence,
By Lemma 4.2(1) and the choices of and , for k sufficiently large, we obtain

where the last inequality follows from .

Therefore,

which implies that (2.3) holds with for all k sufficiently large, i.e., . This completes the proof. □

From now on, we assume that k is large enough so that .

Lemma 4.4Suppose that the assumptions in Lemma  4.2 hold. Set. Then for allksufficiently large, there exists a positive constantsuch that
Proof Set
Then and . Hence, the vectors and are orthogonal. That is,
(4.8)
where is the angle between and . Because and , the angle between and is also . By (4.7), we obtain
which implies that the vectors , and constitute a triangle. Since and . So for all k sufficiently large, we have
which, together with (4.8) and Lemma 4.2(1), we obtain

where . This completes the proof. □

Now, we turn our attention to local rate of convergence analysis.

Theorem 4.1Suppose that the assumptions in Lemma  4.2 hold. Then the sequenceQ-superlinearly converges to 0.

Proof By the definition of , Lemma 4.2(1) and (4.4), for sufficiently large k, we have
which implies that because . Thus, for k sufficiently large, which, together with (4.2), Lemma 4.2, Lemma 4.4, and the definition of , we obtain
Because is bounded, there exists a positive constant such that
(4.9)
On the other hand, from Lemma 3.2, we know that
where S is the solution set of problem (1.1). Since , it follows from Lemma 2.1 that
which implies that
Therefore, together with inequalities (4.1), (4.5), and (4.9), we have

which implies that the order of superlinear convergence is at least 1.5. This completes the proof. □

Remark 4.1 Compared with the proof of the locally superlinear convergence in literatures [15, 16], our conditions are weaker.

## 5 Numerical experiments

In this section, we present some numerical experiments results to show the efficiency of our method. The MATLAB codes are run on a notebook computer with CPU2.10GHZ under MATLAB Version 7.0. Just as done in [16], we take and use the left division operation in MATLAB to solve the system of linear equations (2.1) at each iteration. We choose , , , , and . ‘Iter.’ denotes the number of iteration and ‘CPU’ denotes the CPU time in seconds. We choose as the stop criterion. The example is tested in [16].

Example Let
and the constraint set C be taken as
From Tables 1-2, we can see that our algorithm is efficient if parameters are chosen properly. We can also observe that the algorithm’s operation results change with the value of a. When we take , the operation results are not best, that is to say, the direction is not an optimal one.
Table 1

Numerical results of Example with

Initial point

Iter.

CPU

(3,0,0,0)

11

0.10

1.07 × 10−8

(1,1,0,0)

13

0.09

1.62 × 10−9

(0,1,0,1)

15

0.04

2.46 × 10−9

(0,0,0,1)

21

0.18

9.92 × 10−10

(1,0,0,2)

16

0.54

5.66 × 10−10

Table 2

Numerical results of Example with

Initial point

Iter.

CPU

(3,0,0,0)

11

0.10

1.07 × 10−8

(1,1,0,0)

13

0.12

1.62 × 10−9

(0,1,0,1)

19

0.14

1.17 × 10−9

(0,0,0,1)

18

0.18

1.44 × 10−9

(1,0,0,2)

15

0.21

7.88 × 10−9

## Acknowledgements

The author wish to thank the anonymous referees for their suggestions and comments. This work is also supported by the Educational Science Foundation of Chongqing, Chongqing of China (Grant No. KJ111309).