Some fixed point results regarding convex contractions of Presić type

Abstract

In the present paper, we introduce new types of Presić operators. These operators generalize the well-known Istrăţescu mappings, known as convex contractions. Also, we study the existence and uniqueness of fixed points for this type of operators and the convergence of one-step sequence toward the unique fixed point. Also, data dependence results are presented. Finally, some examples are given, suggesting that the above mappings are proper generalizations of convex contractions of second order

Authors

Cristian Daniel Alecsa
Department of Mathematics Babes-Bolyai University, Cluj-Napoca, Romania
Tiberiu Popoviciu Institute of Numerical Analysis Romanian Academy Cluj-Napoca, Romania

Keywords

Convex contractions; fixed point; Presic operators; data dependence.

Paper coordinates

C.-D. Alecsa, Some fixed point results regarding convex contractions of Presić type, J. Fixed Point Theory Appl., 20 (2018), art. 7,
DOI: 10.1007/s11784-018-0488-7

PDF

About this paper

Journal

Journal of Fixed Point Theory and Applications

Publisher Name

Springer

Print ISSN

1661-7738

Online ISSN

google scholar link

[1] Abbas, M., Ilić, D., Nazir, T. Iterative approximation of fixed points of generalized weak Presic type k-step iterative method for a class of operators. Filomat 29(4), 713–724 (2015)

[2] Alghamdi, M.A., Alnafei, S.H., Radenovic, S., Shahzad, N., Fixed point theorem for convex contraction mappings on cone metric spaces. Math. Comput. Model. 54, 2020–2026 (2011)

[3] Alnafei, S.H., Radenovic, S., Shahzad, N.,  Fixed point theorems for mappings with convex diminishing diameters on cone metric spaces. Appl. Math. Lett. 24, 2162–2166 (2011)

[4] Banach, S., Sur les opérations dans les ensembles abstraits et leur application aux equations integrales. Fund. Math. 3, 133–181 (1922)

[5] Berinde, V., Păcurar, M., Stability of k-step fixed point iterative methods for some Pres̆ić type contractive mappings. J. Inequal. Appl. 2014:149 (2014)

[6] Ćirić, L.B., Pres̆ić, L.B.,  On Presic type generalisation of Banach contraction mapping principle. Acta Math. Univ. Comenian 76(2), 143–147 (2007)

[7] Istrăţescu, V., Some fixed point theorems for convex contraction mappings and convex nonexpansive mappings (I). Lib. Math. 1, 151–164 (1981)

[8] Istrăţescu, V., Some fixed point theorems for convex contraction mappings and mappings with convex diminishing diameters – I. Ann. Mat. Pura Appl. 130, 89–104 (1982)

[9] Istrăţescu, V., Some fixed point theorems for convex contraction mappings and mappings with convex diminishing diameters. II. Ann. Mat. Pura Appl. 134, 327–362 (1983)

[10] Khan, M.S., Berzig, M., Samet, B., Some convergence results for iterative sequences of Pres̆ić type and applications. Adv. Differ. Equ. 38 (2012). https://doi.org/10.1186/1687-1847-2012-38

[11] Miculescu, R., Mihail, A., A generalization of Matkowski’s fixed point theorem and Istrăţescu’s fixed point theorem concerning convex contractions. J. Fixed Point Theory Appl. 19, 1525–1533 (2017)

[12] Mureşan, V., Mureşan, A.S., On the theory of fixed point theorems for convex contraction mappings. Carpatian J. Math. 31(3), 365–371 (2015)

[13] Păcurar, M.,  Approximating common fixed points of Presić–Kannan type operators by a multi-step iterative method. An. Şt. Univ. Ovidiu Constanţa 17(1), 153–168 (2009)

[14] Păcurar, M.,  A multi-step method for approximating common fixed points of Presić–Rus type operators on metric spaces. Studia Univ. Babeş-Bolyai Math. LV(1), 149–162 (2010)

[15] Păcurar, M.,  Fixed points of almost Presić operators by a k-step iterative method. An. Ştiinţ. Univ. AI.I. Cuza Iaşi Mat. 57:199–210 (2011)

[16] Pathak, H.K., George, R., Nabwey, H.A., El-Paoumy, M.S., Reshma, K.P., Some generalized fixed point results in b-metric space and application to matrix equations. Fixed Point Theory Appl. 2015:101 (2015)

[17] Pres̆ić, S.B., Sur une classe d’inéquations aux differences finies et sur la convergence de certain suites. Pub. de l’Inst. Math. Belgrade 5(19), 75–78 (1965)

[18] Rus, I., An iterative method for the solution of the equation x=f(x,…,x). Anal. Numér. Thor. Approx. 10(1), 95–100 (1981)

[19] Sastry, K.P.R., Rao, ChS, Sekhar, C., Balaiah, M., A fixed point theorem for cone convex contractions of order m≥2. Int. J. Math. Sci. Eng. Appl. 6(1), 263–271 (2012)

[20] Shukla, S., Radenovic, S., Presić–Maya type theorems in ordered metric spaces. Gulf J. Math. 2(2), 73–82 (2014)

[21] Shukla, S., Radenovic, S., Pantelić, S., Some fixed point theorems for Pres̆ić–Hardy–Rogers type contractions in metric spaces. J. Math. (2013). https://doi.org/10.1155/2013/295093

[22] Shukla, S., Radenovic, S., Some generalizations of Pres̆ić type mappings and applications. An. Ştiinţ. Univ. AI.I. Cuza Iaşi Mat. (2015). https://doi.org/10.1515/aicu-2015-0026

Paper (preprint) in HTML form

Some fixed point results regarding convex contractions of Presić type

Cristian Daniel Alecsa
Abstract

In the present paper, we introduce new types of Presić operators. These operators generalize the well-known Istrăţescu mappings, known as convex contractions. Also, we study the existence and uniqueness of fixed points for this type of operators and the convergence of one-step sequence toward the unique fixed point. Also, data dependence results are presented. Finally, some examples are given, suggesting that the above mappings are proper generalizations of convex contractions of second order.

Mathematics Subject Classification. Primary 47H10 ; Secondary 54H25.
Keywords. Convex contractions, fixed point, Presić operators, data dependence.

1. Introduction and preliminaries

The main starting point of the present paper is the well-known contraction principle, given by Stefan Banach [4]. A generalization related to the Banach contraction principle was given by Presić [17]. He proved the following theorem.

Theorem 1.1. Let ( X,dX,d ) be a complete metric space and f:XkXf:X^{k}\rightarrow X a mapping satisfying the following condition:

d(f(x0,,xk1),f(x1,,xk))i=0k1αid(xi,xi+1)d\left(f\left(x_{0},\ldots,x_{k-1}\right),f\left(x_{1},\ldots,x_{k}\right)\right)\leq\sum_{i=0}^{k-1}\alpha_{i}d\left(x_{i},x_{i+1}\right) (1.1)

for each x0,,xkXx_{0},\ldots,x_{k}\in X, where α0++αk1[0,1)\alpha_{0}+\cdots+\alpha_{k-1}\in[0,1).
Then, there exists a unique point xXx^{*}\in X, such that f(x,x,,x)=xf\left(x^{*},x^{*},\ldots,x^{*}\right)=x^{*}. Moreover, the kk-step sequence defined by

xn+k=f(xn,,xn+k1)x_{n+k}=f\left(x_{n},\ldots,x_{n+k-1}\right) (1.2)

is convergent and satisfies limnxn=f(limnxn,,limnxn)\lim_{n\rightarrow\infty}x_{n}=f\left(\lim_{n\rightarrow\infty}x_{n},\ldots,\lim_{n\rightarrow\infty}x_{n}\right).

The kk-step iterative sequence given by (1.2) can be regarded as a nonlinear difference equation. Furthermore, if the sequence (xn)n\left(x_{n}\right)_{n\in\mathbb{N}} is convergent, then its limit is a fixed point for the mapping ff.
Many authors have generalized the contractive condition given by Presić. One of the earliest generalization was made by Presić and Ćirić in [6]. We recall their results.

Theorem 1.2. Let ( X,dX,d ) be a complete metric space. Consider a mapping f:XkXf:X^{k}\rightarrow X satisfying the following contractive-type condition

d(f(x0,,xk1),f(x1,,xk))λmax{d(xi,xi+1):0ik1}d\left(f\left(x_{0},\ldots,x_{k-1}\right),f\left(x_{1},\ldots,x_{k}\right)\right)\leq\lambda\max\left\{d\left(x_{i},x_{i+1}\right):0\leq i\leq k-1\right\} (1.3)

where λ(0,1)\lambda\in(0,1) and x0,,xkx_{0},\ldots,x_{k} are arbitrary given elements of XX.
Then, there exists a point xXx^{*}\in X for which f(x,,x)=xf\left(x^{*},\ldots,x^{*}\right)=x^{*}.
Furthermore, if x0,,xk1x_{0},\ldots,x_{k-1} are arbitrary elements from XX, then the sequence (xn)n\left(x_{n}\right)_{n\in\mathbb{N}} defined by

xn+k=f(xn,,xn+k1)x_{n+k}=f\left(x_{n},\ldots,x_{n+k-1}\right) (1.4)

is convergent and satisfies limnxn=f(limnxn,limnxn)\lim_{n\rightarrow\infty}x_{n}=f\left(\lim_{n\rightarrow\infty}x_{n},\ldots\lim_{n\rightarrow\infty}x_{n}\right). Additionally, if on the diagonal ΔXk\Delta\subset X^{k}, we have that

d(f(u,,u),f(v,,v))<d(u,v),d(f(u,\ldots,u),f(v,\ldots,v))<d(u,v), (1.5)

for each u,vXu,v\in X, with uvu\neq v, then xx^{*} is the unique point in XX.
Also, in 1981, Rus [18] generalized (Theorem 1.1) and proved the existence and uniqueness of a fixed point in XX, i.e., a point xx^{*} that satisfies x=f(x,,x)x^{*}=f\left(x^{*},\ldots,x^{*}\right), for a mapping f:XkXf:X^{k}\rightarrow X, using the following condition:

d(f(x0,,xk1),f(x1,,xk))φ(d(x0,x1),,d(xk1,xk))d\left(f\left(x_{0},\ldots,x_{k-1}\right),f\left(x_{1},\ldots,x_{k}\right)\right)\leq\varphi\left(d\left(x_{0},x_{1}\right),\ldots,d\left(x_{k-1},x_{k}\right)\right) (1.6)

where the continuous function φ\varphi is such that

{φ(r)φ(s), where rs, with r,s+k,φ(t,,t)<t, for each t+,i=0φi(r)<+, for each r+k,φ(t,0,,0)+φ(0,t,0,,0)++φ(0,,0,t)φ(t,,t),t+.\left\{\begin{array}[]{l}\varphi(r)\leq\varphi(s),\text{ where }r\leq s,\text{ with }r,s\in\mathbb{R}_{+}^{k},\\ \varphi(t,\ldots,t)<t,\text{ for each }t\in\mathbb{R}_{+},\\ \sum_{i=0}^{\infty}\varphi^{i}(r)<+\infty,\text{ for each }r\in\mathbb{R}_{+}^{k},\\ \varphi(t,0,\ldots,0)+\varphi(0,t,0,\ldots,0)+\cdots+\varphi(0,\ldots,0,t)\leq\varphi(t,\ldots,t),\forall t\in\mathbb{R}_{+}.\end{array}\right.

More recently, Păcurar [14] gave a generalization of Rus’s result and studied the existence of coincidence and common fixed points for a pair of mappings (f,g)(f,g), where f:XkXf:X^{k}\rightarrow X and g:XXg:X\rightarrow X, such that
d(f(x0,,xk1),f(x1,,xk))φ(d(g(x0),g(x1)),,d(g(xk1),g(xk)))d\left(f\left(x_{0},\ldots,x_{k-1}\right),f\left(x_{1},\ldots,x_{k}\right)\right)\leq\varphi\left(d\left(g\left(x_{0}\right),g\left(x_{1}\right)\right),\ldots,d\left(g\left(x_{k-1}\right),g\left(x_{k}\right)\right)\right),
where φ:+k+k\varphi:\mathbb{R}_{+}^{k}\rightarrow\mathbb{R}_{+}^{k} is endowed with the properties presented above.
In two further articles [13] and [15], Păcurar studied fixed points and common fixed points of Presić-type operators under the following contractive-type conditions d(f(x0,,xk1),f(x1,,xk))ai=0kd(g(xi),f(xi,,xi))d\left(f\left(x_{0},\ldots,x_{k-1}\right),f\left(x_{1},\ldots,x_{k}\right)\right)\leq a\sum_{i=0}^{k}d\left(g\left(x_{i}\right),f\left(x_{i},\ldots,x_{i}\right)\right) for a pair ( f,gf,g ), respectively :

d(f(x0,,xk1),f(x1,,xk))i=1kδid(xi1,xi)+M¯(x0,xk)d\left(f\left(x_{0},\ldots,x_{k-1}\right),f\left(x_{1},\ldots,x_{k}\right)\right)\leq\sum_{i=1}^{k}\delta_{i}d\left(x_{i-1},x_{i}\right)+\bar{M}\left(x_{0},x_{k}\right)

where M¯(x0,xk)=Lmin{d(x0,f(x0,,x0)),d(xk,f(xk,,xk))\bar{M}\left(x_{0},x_{k}\right)=L\min\left\{d\left(x_{0},f\left(x_{0},\ldots,x_{0}\right)\right),d\left(x_{k},f\left(x_{k},\ldots,x_{k}\right)\right)\right., d(x0,f(xk,,xk)),d(xk,f(x0,,x0)),d(xk,f(x0,,xk1))}\left.d\left(x_{0},f\left(x_{k},\ldots,x_{k}\right)\right),d\left(x_{k},f\left(x_{0},\ldots,x_{0}\right)\right),d\left(x_{k},f\left(x_{0},\ldots,x_{k-1}\right)\right)\right\}, with L>L> 0 . Other generalizations of Presić-type mappings were made by Shukla and Radenovic. They study existence and uniqueness for fixed points of various mappings in complete metric spaces and in complete ordered metric spaces, [20,21] and [22]. The most general condition, called Presić-Hardy-Rogers, has the following form:

d(f(x0,,xk1),f(x1,,xk))\displaystyle d\left(f\left(x_{0},\ldots,x_{k-1}\right),f\left(x_{1},\ldots,x_{k}\right)\right)
i=0k1d(xi,xi+1)+i=0kj=0kβi,jd(xi,f(xj,,xj)).\displaystyle\quad\leq\sum_{i=0}^{k-1}d\left(x_{i},x_{i+1}\right)+\sum_{i=0}^{k}\sum_{j=0}^{k}\beta_{i,j}d\left(x_{i},f\left(x_{j},\ldots,x_{j}\right)\right). (1.8)

Also, existence and uniqueness of fixed points for various Presić type mappings were studied. We let the reader follow [1,10][1,10] and [16]. Furthermore, we remind that the mentioned authors gave examples in metric spaces and b-metric spaces. Also, in [16] applications to matrix equations were given. Finally, since our aim is to study the existence and uniqueness of fixed points for some new types of mappings f:XkXf:X^{k}\rightarrow X, for more research papers regarding Presić-type mappings, we let the reader follow [5,13,15] and [17]. Since we shall extend the concept of convex contractions to Presić operators, we remind the definition of convex contraction of second order, given by Istrăţescu in [8].

Definition 1.3. Let ( X,dX,d ) be a metric space. Consider a continuous mapping f:XXf:X\rightarrow X. f is said to be a convex contraction of order 2 if there exists a,b(0,1)a,b\in(0,1), such that for each x,yXx,y\in X

d(f2(x),f2(y))ad(f(x),f(y))+bd(x,y)d\left(f^{2}(x),f^{2}(y)\right)\leq ad(f(x),f(y))+bd(x,y) (1.9)

where a+b<1a+b<1.
Furthermore, in the same paper, Istrăţescu introduced convex contractions of order n , like follows.

Definition 1.4. Let ( X,dX,d ) be a metric space. Consider a continuous mapping f:XXf:X\rightarrow X. f is said to be a convex contraction of order nn if there exists a0,,an1(0,1)a_{0},\ldots,a_{n-1}\in(0,1), such that for each x,yXx,y\in X

d(fn(x),fn(y))a0d(x,y)+a1d(f(x),f(y))+\displaystyle d\left(f^{n}(x),f^{n}(y)\right)\leq a_{0}d(x,y)+a_{1}d(f(x),f(y))+\cdots
+an1d(fn1(x),fn1(y)), where a0++an1<1\displaystyle\quad+a_{n-1}d\left(f^{n-1}(x),f^{n-1}(y)\right),\text{ where }a_{0}+\cdots+a_{n-1}<1 (1.10)

Even though in the definitions given by Istrăţescu, in the case of convex contractions of order 2 , the coefficients a,ba,b are in ( 0,1 ) and in the case of convex contractions of order n2n\geq 2, the coefficients a0,,an1a_{0},\ldots,a_{n-1} lie in the interval (0,1)(0,1), we shall employ the fact that the coefficients can be in [0,1)[0,1) as in [19]. This change will be very useful for the examples given in the last section.

Also, Istrăţescu studied other types of continuous operators in [7] and [9] and Sastry et.al. [19] studied the existence and uniqueness principles for convex contractions of order m2m\geq 2. Moreover, we remind that Mureşan and Mureşan [12] gave theorems regarding data dependence and qualitative properties for convex contractions of order 2.
Finally, other authors have studied qualitative properties and developed existence and uniqueness theorems for convex contractions of order 2 and for other types of operators, such as convex contractions with diminishing diameters. We let the reader follow [2,3][2,3] and [11].

2. Main results

In this section, we introduce new types of Presić operators that will generalize convex contractions of order 2, namely Presić convex contraction of the first kind and Presić convex contraction of the second kind. Furthermore, data dependence results are also given and some important remarks concerning our new operators are made in order to emphasize the fact that our mappings generalize contractions when k=1k=1, respectively Presić contractions when k1k\geq 1.

Definition 2.1. Let ( X,dX,d ) be a metric space. Let x0,x1,,xk1,xkx_{0},x_{1},\ldots,x_{k-1},x_{k} be arbitrary elements from XX. Consider the coefficients αi[0,1)\alpha_{i}\in[0,1), with i=0,k1¯i=\overline{0,k-1} and βij[0,1)\beta_{ij}\in[0,1), with i,j=0,k¯i,j=\overline{0,k}.
A mapping f:XkXf:X^{k}\rightarrow X satisfying

d(f(f(x0,,x0),,f(xk1,,xk1)),f(f(x1,,x1),,f(xk,,xk)))\displaystyle d\left(f\left(f\left(x_{0},\ldots,x_{0}\right),\ldots,f\left(x_{k-1},\ldots,x_{k-1}\right)\right),f\left(f\left(x_{1},\ldots,x_{1}\right),\ldots,f\left(x_{k},\ldots,x_{k}\right)\right)\right)
i=0k1αid(xi,xi+1)+i,j=0kβijd(f(xi,,xi),f(xj,,xj))\displaystyle\quad\leq\sum_{i=0}^{k-1}\alpha_{i}d\left(x_{i},x_{i+1}\right)+\sum_{i,j=0}^{k}\beta_{ij}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{j},\ldots,x_{j}\right)\right)

is called a Presić convex contraction of the first kind.
Definition 2.2. Let ( X,dX,d ) be a metric space. Let x0,x1,,xk1,xkx_{0},x_{1},\ldots,x_{k-1},x_{k} be arbitrary elements from XX. Consider the coefficients αi[0,1)\alpha_{i}\in[0,1), with i=0,k1¯i=\overline{0,k-1} and βij[0,1)\beta_{ij}\in[0,1), with i,j=0,k¯i,j=\overline{0,k}.
A mapping f:XkXf:X^{k}\rightarrow X satisfying

d(f(f(x0,,xk1),,f(x0,,xk1)),f(f(x1,,xk),,f(x1,,xk)))\displaystyle d\left(f\left(f\left(x_{0},\ldots,x_{k-1}\right),\ldots,f\left(x_{0},\ldots,x_{k-1}\right)\right),f\left(f\left(x_{1},\ldots,x_{k}\right),\ldots,f\left(x_{1},\ldots,x_{k}\right)\right)\right)
i=0k1αid(xi,xi+1)+i,j=0kβijd(f(xi,,xi),f(xj,,xj))\displaystyle\leq\sum_{i=0}^{k-1}\alpha_{i}d\left(x_{i},x_{i+1}\right)+\sum_{i,j=0}^{k}\beta_{ij}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{j},\ldots,x_{j}\right)\right)

is called a Presić convex contraction of the second kind.
Definition 2.3. Let ( X,dX,d ) be a metric space and f:XkXf:X^{k}\rightarrow X a mapping. Then, the operator F:XXF:X\rightarrow X, defined as F(x)=f(x,,x)F(x)=f(x,\ldots,x), for each xXx\in X is called the associated operator of ff.

The first theorem that we present is related to the existence and uniqueness of the fixed point of ff, i.e., the element xXx^{*}\in X that satisfies x=f(x,,x)x^{*}=f\left(x^{*},\ldots,x^{*}\right). More precisely, we shall present a theorem involving the
fixed point of the associated operator FF, i.e., xXx^{*}\in X such that x=F(x)x^{*}=F\left(x^{*}\right) of the Presić convex contractions of the first kind, respectively, for the fixed point of the associated operator of Presić convex contractions of the second kind.

Theorem 2.4. Let ( X,dX,d ) be a complete metric space.
(i) Let f:XkXf:X^{k}\rightarrow X a Presić convex contraction of the first kind which is a continuous mapping. Suppose that the coefficients of the mapping ff from (Definition 1.3) satisfy

i=0k1αi+2p=1k(i=0kpj=kp+1kβij)(0,1).\sum_{i=0}^{k-1}\alpha_{i}+2\sum_{p=1}^{k}\left(\sum_{i=0}^{k-p}\sum_{j=k-p+1}^{k}\beta_{ij}\right)\in(0,1). (2.1)

Then, ff has a unique fixed point xx^{*} and the sequence (xn)n\left(x_{n}\right)_{n\in\mathbb{N}} defined as xn+1=f(xn,,xn)x_{n+1}=f\left(x_{n},\ldots,x_{n}\right), for each n1n\geq 1, is convergent to xx^{*}.
(ii) Consider f:XkXf:X^{k}\rightarrow X a Presić convex contraction of the second kind that is a continuous mapping. Suppose that the coefficients from the (Definition 1.4) satisfy the same condition as before, namely (2.1).
Then, as in the previous case, ff has a unique fixed point xx^{*} and the sequence (xn)n\left(x_{n}\right)_{n\in\mathbb{N}} defined as xn+1=f(xn,,xn)x_{n+1}=f\left(x_{n},\ldots,x_{n}\right), for each n1n\geq 1, is convergent to xx^{*}

Proof. (i) Let f:XkXf:X^{k}\rightarrow X be a continuous Presić convex contraction of the first kind.
First, we consider the associated operator FF of ff, i.e., :F:XX:F:X\rightarrow X, defined as F(x)=f(x,,x)F(x)=f(x,\ldots,x). We show that the self-mapping FF is a convex contraction of the second order. Furthermore, since f:XkXf:X^{k}\rightarrow X is continuous, then F:XXF:X\rightarrow X is also a continuous mapping.
From now on, we make the remark that the technique used in the present proof is based on [13,15] and [21].
Let x,yXx,y\in X be arbitrary elements from the metric space XX. We apply the triangle inequality by k times.

D:=\displaystyle D= d(F2x,F2y)=d(F(Fx),F(Fy))=d(f(Fx,,Fx),f(Fy,,Fy))\displaystyle d\left(F^{2}x,F^{2}y\right)=d(F(Fx),F(Fy))=d(f(Fx,\ldots,Fx),f(Fy,\ldots,Fy))
\displaystyle\leq d(f(Fx,Fx,,Fx,Fx),f(Fx,Fx,,Fx,Fy))\displaystyle d(f(Fx,Fx,\ldots,Fx,Fx),f(Fx,Fx,\ldots,Fx,Fy))
+d(f(Fx,Fx,,Fx,Fy),f(Fx,Fx,,Fy,Fy))\displaystyle+d(f(Fx,Fx,\ldots,Fx,Fy),f(Fx,Fx,\ldots,Fy,Fy))
\displaystyle\cdots
+d(f(Fx,Fy,,Fy,Fy),f(Fy,Fy,,Fy,Fy)).\displaystyle+d(f(Fx,Fy,\ldots,Fy,Fy),f(Fy,Fy,\ldots,Fy,Fy)).

Now, let us denote the distances from the right hand side by D1,,Dk1,DkD_{1},\ldots,D_{k-1},D_{k}.
Also, from now on, let us denote by Dxx=d(Fx,Fx),Dyy=d(Fy,Fy)D_{xx}=d(Fx,Fx),D_{yy}=d(Fy,Fy), Dxy=d(Fx,Fy)D_{xy}=d(Fx,Fy), and Dyx=d(Fy,Fx)D_{yx}=d(Fy,Fx). Our aim is to compute each of Di,i=1,k¯D_{i},i=\overline{1,k}. Furthermore, we make the remark that since DxxD_{xx} and DyyD_{yy} are equal to 0 , then the coefficients βij\beta_{ij} which shall appear at DxxD_{xx} and DyyD_{yy} in the computation of the distances DiD_{i} are omitted.

D1=\displaystyle D_{1}= d(f(Fx,Fx,,Fx,Fx),f(Fx,Fx,,Fx,Fy))\displaystyle d(f(Fx,Fx,\ldots,Fx,Fx),f(Fx,Fx,\ldots,Fx,Fy))
=\displaystyle= d(f(f(x,,x),f(x,,x),,f(x,,x)),\displaystyle d(f(f(x,\ldots,x),f(x,\ldots,x),\ldots,f(x,\ldots,x)),
f(f(x,,x),,f(x,,x),f(y,,y))).\displaystyle f(f(x,\ldots,x),\ldots,f(x,\ldots,x),f(y,\ldots,y))).

Applying the contractive-type condition with x0=x1==xk1=xx_{0}=x_{1}=\cdots=x_{k-1}=x and xk=yx_{k}=y, we get that

D1\displaystyle D_{1}\leq [α0d(x,x)+α1d(x,x)++αk2d(x,x)+αk1d(x,y)]\displaystyle{\left[\alpha_{0}d(x,x)+\alpha_{1}d(x,x)+\cdots+\alpha_{k-2}d(x,x)+\alpha_{k-1}d(x,y)\right]}
+i,j=0kβijd(f(xi,,xi),f(xj,,xj))=αk1d(x,y)\displaystyle+\sum_{i,j=0}^{k}\beta_{ij}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{j},\ldots,x_{j}\right)\right)=\alpha_{k-1}d(x,y)
+i,j=0kβijd(f(xi,,xi),f(xj,,xj))\displaystyle+\sum_{i,j=0}^{k}\beta_{ij}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{j},\ldots,x_{j}\right)\right)
=\displaystyle= αk1d(x,y)+i,j=0kβijd(Fxi,Fxj).\displaystyle\alpha_{k-1}d(x,y)+\sum_{i,j=0}^{k}\beta_{ij}d\left(Fx_{i},Fx_{j}\right).

Now, since the terms from the contractive-type condition x0,,xk1,xkx_{0},\ldots,x_{k-1},x_{k} take values in {x,y}\{x,y\}, for simplifications we divide the coefficients of the sum of the right hand side as follows.
For DxyD_{xy}, we have that j{k}j\in\{k\} and i0,k1¯i\in\overline{0,k-1}, so we get the coefficient of Dx,yD_{x,y}, i.e., i=0k1βi,k\sum_{i=0}^{k-1}\beta_{i,k}.
In a similar way, for DyxD_{yx}, we have that i{k}i\in\{k\} and j0,k1¯j\in\overline{0,k-1}, so the coefficient of DyxD_{yx} is j=0k1βkj\sum_{j=0}^{k-1}\beta_{kj}.
All of these imply that

D1αk1d(x,y)+[i=0k1βik]Dxy+[j=0k1βkj]Dyx.D_{1}\leq\alpha_{k-1}d(x,y)+\left[\sum_{i=0}^{k-1}\beta_{ik}\right]D_{xy}+\left[\sum_{j=0}^{k-1}\beta_{kj}\right]D_{yx}.

Now, for D2D_{2}, we have, in a similar way, the following:

D2=\displaystyle D_{2}= d(f(Fx,Fx,,Fx,Fy),f(Fx,Fx,,Fy,Fy))\displaystyle d(f(Fx,Fx,\ldots,Fx,Fy),f(Fx,Fx,\ldots,Fy,Fy))
=\displaystyle= d(f(f(x,,x),f(x,,x),,f(x,,x),f(y,,y)),\displaystyle d(f(f(x,\ldots,x),f(x,\ldots,x),\ldots,f(x,\ldots,x),f(y,\ldots,y)),
f(f(x,,x),,f(x,,x),f(y,,y),f(y,,y))).\displaystyle f(f(x,\ldots,x),\ldots,f(x,\ldots,x),f(y,\ldots,y),f(y,\ldots,y))).

Using a similar approach as for D1D_{1}, applying the contractive-type condition with x0=x1==xk2=xx_{0}=x_{1}=\cdots=x_{k-2}=x and xk1=xk=yx_{k-1}=x_{k}=y, we get that
D2αk2d(x,y)+[i=0k2βi,k1+i=0k2βik]Dxy+[j=0k2βk1,j+j=0k2βkj]DyxD_{2}\leq\alpha_{k-2}d(x,y)+\left[\sum_{i=0}^{k-2}\beta_{i,k-1}+\sum_{i=0}^{k-2}\beta_{ik}\right]D_{xy}+\left[\sum_{j=0}^{k-2}\beta_{k-1,j}+\sum_{j=0}^{k-2}\beta_{kj}\right]D_{yx}.
The reasoning is the same for every DiD_{i}, where i=1,k¯i=\overline{1,k}. For the sake of convenience, the last distance DkD_{k} has the following form.

Dk=\displaystyle D_{k}= d(f(Fx,Fy,,Fy,Fy),f(Fy,Fy,,Fy,Fy))\displaystyle d(f(Fx,Fy,\ldots,Fy,Fy),f(Fy,Fy,\ldots,Fy,Fy))
=\displaystyle= d(f(f(x,,x),f(y,,y),,f(y,,y),f(y,,y)),\displaystyle d(f(f(x,\ldots,x),f(y,\ldots,y),\ldots,f(y,\ldots,y),f(y,\ldots,y)),
f(f(y,,y),,f(y,,y),f(y,,y))).\displaystyle f(f(y,\ldots,y),\ldots,f(y,\ldots,y),f(y,\ldots,y))).

We apply the contractive-type condition with x0=xx_{0}=x and x1==xk1=xk=yx_{1}=\cdots=x_{k-1}=x_{k}=y. So, it follows that

Dk\displaystyle D_{k}\leq [α0d(x,y)+α1d(y,y)++αk2d(y,y)+αk1d(y,y)]\displaystyle{\left[\alpha_{0}d(x,y)+\alpha_{1}d(y,y)+\cdots+\alpha_{k-2}d(y,y)+\alpha_{k-1}d(y,y)\right]}
+i,j=0kβijd(Fxi,Fxj)\displaystyle+\sum_{i,j=0}^{k}\beta_{ij}d\left(Fx_{i},Fx_{j}\right)
=\displaystyle= α0d(x,y)+i,j=0kβijd(f(xi,,xi),f(xj,,xj))α0d(x,y)\displaystyle\alpha_{0}d(x,y)+\sum_{i,j=0}^{k}\beta_{ij}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{j},\ldots,x_{j}\right)\right)\alpha_{0}d(x,y)
+i,j=0kβijd(Fxi,Fxj)\displaystyle+\sum_{i,j=0}^{k}\beta_{ij}d\left(Fx_{i},Fx_{j}\right)

Now, since x0,,xk1,xkx_{0},\ldots,x_{k-1},x_{k} take values in {x,y}\{x,y\}, we have that the coefficients Dxy,DyxD_{xy},D_{yx} of DkD_{k} are given as follows.
For Dxy=d(Fx,Fy)=d(Fxi,Fxj)D_{xy}=d(Fx,Fy)=d\left(Fx_{i},Fx_{j}\right), we have that i{0}i\in\{0\} and j1,k¯j\in\overline{1,k}, so we get the coefficient of Dx,yD_{x,y} is j=1kβ0,j\sum_{j=1}^{k}\beta_{0,j}.
Also, for DyxD_{yx}, we have that i{1,,k}i\in\{1,\ldots,k\} and j{0}j\in\{0\}, so the coefficient of DyxD_{yx} is given by i=1kβi,0\sum_{i=1}^{k}\beta_{i,0}.
From the above values of the coefficients DxyD_{xy}, and DyxD_{yx}, we get that Dkα0d(x,y)+[j=1kβ0j]Dxy+[i=1kβi0]DyxD_{k}\leq\alpha_{0}d(x,y)+\left[\sum_{j=1}^{k}\beta_{0j}\right]D_{xy}+\left[\sum_{i=1}^{k}\beta_{i0}\right]D_{yx}. Now, we simplify the values of DiD_{i}, for each i=1,k¯i=\overline{1,k}, such as :

D1αk1d(x,y)+[i=0k1βik]Dxy+[j=0k1βkj]Dyx\displaystyle D_{1}\leq\alpha_{k-1}d(x,y)+\left[\sum_{i=0}^{k-1}\beta_{ik}\right]D_{xy}+\left[\sum_{j=0}^{k-1}\beta_{kj}\right]D_{yx}
D2αk2d(x,y)+[i=0k2j=k1kβij]Dxy+[i=k1kj=0k2βij]Dyx\displaystyle D_{2}\leq\alpha_{k-2}d(x,y)+\left[\sum_{i=0}^{k-2}\sum_{j=k-1}^{k}\beta_{ij}\right]D_{xy}+\left[\sum_{i=k-1}^{k}\sum_{j=0}^{k-2}\beta_{ij}\right]D_{yx}
\displaystyle\ldots
Dkα0d(x,y)+[j=1kβ0j]Dxy+[i=1kβi0]Dyx\displaystyle D_{k}\leq\alpha_{0}d(x,y)+\left[\sum_{j=1}^{k}\beta_{0j}\right]D_{xy}+\left[\sum_{i=1}^{k}\beta_{i0}\right]D_{yx}

So, for DD, it follows that
DAd(x,y)+CDxy+EDyxD\leq Ad(x,y)+C\cdot D_{xy}+E\cdot D_{yx}, where the sums A,C\mathrm{A},\mathrm{C} and E have the following form :

A\displaystyle A :=i=0k1αi\displaystyle=\sum_{i=0}^{k-1}\alpha_{i}
C\displaystyle C :=[i=0k1βik+i=0k2j=k1kβij++j=1kβ0j]\displaystyle=\left[\sum_{i=0}^{k-1}\beta_{ik}+\sum_{i=0}^{k-2}\sum_{j=k-1}^{k}\beta_{ij}+\cdots+\sum_{j=1}^{k}\beta_{0j}\right]
E\displaystyle E :=[j=0k1βkj+i=k1kj=0k2βij++i=2kj=01βij+i=1kβi0].\displaystyle=\left[\sum_{j=0}^{k-1}\beta_{kj}+\sum_{i=k-1}^{k}\sum_{j=0}^{k-2}\beta_{ij}+\cdots+\sum_{i=2}^{k}\sum_{j=0}^{1}\beta_{ij}+\sum_{i=1}^{k}\beta_{i0}\right].

Moreover, we mention additional forms of some terms in CC, and EE, i.e.,

i=0k1βik=i=0k1j=kkβij,j=1kβ0j=i=00j=1kβij\displaystyle\sum_{i=0}^{k-1}\beta_{ik}=\sum_{i=0}^{k-1}\sum_{j=k}^{k}\beta_{ij},\sum_{j=1}^{k}\beta_{0j}=\sum_{i=0}^{0}\sum_{j=1}^{k}\beta_{ij}
j=0k1βkj=i=kkj=0k1βij,i=1kβi0=i=1kj=00βij\displaystyle\sum_{j=0}^{k-1}\beta_{kj}=\sum_{i=k}^{k}\sum_{j=0}^{k-1}\beta_{ij},\sum_{i=1}^{k}\beta_{i0}=\sum_{i=1}^{k}\sum_{j=0}^{0}\beta_{ij}

We have that DAd(x,y)+CDxy+EDyxD\leq Ad(x,y)+CD_{xy}+ED_{yx}.
Let’s impose the following condition that does not affect the contractive-type condition, i.e., for each i,j{0,,k},βij=βjii,j\in\{0,\ldots,k\},\beta_{ij}=\beta_{ji}, because βij\beta_{ij} appear as a coefficient to d(Fxi,Fxj)d\left(Fx_{i},Fx_{j}\right) and βji\beta_{ji} appear as a coefficient of the distance d(Fxj,Fxi)d\left(Fx_{j},Fx_{i}\right), which in examples they are equal.
Now, for CC and EE, using the above remark, we infer another observation. In the case of EE, it follows that the first term is

j=0k1βkj=i=kkj=0k1βij=j=0k1i=kkβij=j=0k1i=kkβji.\sum_{j=0}^{k-1}\beta_{kj}=\sum_{i=k}^{k}\sum_{j=0}^{k-1}\beta_{ij}=\sum_{j=0}^{k-1}\sum_{i=k}^{k}\beta_{ij}=\sum_{j=0}^{k-1}\sum_{i=k}^{k}\beta_{ji}.

Changing the notation of index ii with the notation of index jj, we get that the term j=0k1βkj\sum_{j=0}^{k-1}\beta_{kj} is i=0k1j=kkβij=i=0k1βik\sum_{i=0}^{k-1}\sum_{j=k}^{k}\beta_{ij}=\sum_{i=0}^{k-1}\beta_{ik} that is the first term of C .
Now, the second term of EE is

i=k1kj=0k2βij=j=0k2i=k1kβij.\sum_{i=k-1}^{k}\sum_{j=0}^{k-2}\beta_{ij}=\sum_{j=0}^{k-2}\sum_{i=k-1}^{k}\beta_{ij}.

Changing again the notation of the index ii with that of jj, it follows that i=k1kj=0k2βij\sum_{i=k-1}^{k}\sum_{j=0}^{k-2}\beta_{ij} is in fact i=0k2j=k1kβji\sum_{i=0}^{k-2}\sum_{j=k-1}^{k}\beta_{ji} and by symmetry of sums and of βij\beta_{ij} is i=0k2j=k1kβij\sum_{i=0}^{k-2}\sum_{j=k-1}^{k}\beta_{ij}, i.e., the second term of C .
Now, for the last term of EE, we get that

i=1kβi0=i=kkj=00βij=j=00i=1kβij.\sum_{i=1}^{k}\beta_{i0}=\sum_{i=k}^{k}\sum_{j=0}^{0}\beta_{ij}=\sum_{j=0}^{0}\sum_{i=1}^{k}\beta_{ij}.

Changing again the notation of ii with jj, the last term is i=00j=1kβji\sum_{i=0}^{0}\sum_{j=1}^{k}\beta_{ji}, which again by symmetry it is : i=00j=1kβij=j=1kβ0j\sum_{i=0}^{0}\sum_{j=1}^{k}\beta_{ij}=\sum_{j=1}^{k}\beta_{0j}.
So, from the above reasoning we have that the sum CC is in fact EE, so DAd(x,y)+2CDxyD\leq Ad(x,y)+2CD_{xy}.
Moreover, we estimate the sum CC.

C=i=0k1j=kkβij+i=0k2j=k1kβij++i=01j=2kβij+i=00j=1kβij.C=\sum_{i=0}^{k-1}\sum_{j=k}^{k}\beta_{ij}+\sum_{i=0}^{k-2}\sum_{j=k-1}^{k}\beta_{ij}+\cdots+\sum_{i=0}^{1}\sum_{j=2}^{k}\beta_{ij}+\sum_{i=0}^{0}\sum_{j=1}^{k}\beta_{ij}.

Furthermore, it is easy to see that CC can be written as:

C=p=1k(i=0kpj=kp+1kβij).C=\sum_{p=1}^{k}\left(\sum_{i=0}^{k-p}\sum_{j=k-p+1}^{k}\beta_{ij}\right).

Now, since we have shown that D=d(F2x,F2y)Ad(x,y)+2Cd(Fx,Fy)D=d\left(F^{2}x,F^{2}y\right)\leq Ad(x,y)+2Cd(Fx,Fy), from the hypothesis assumption that i=0k1αi+2p=1k(i=0kpj=kp+1kβij)(0,1)\sum_{i=0}^{k-1}\alpha_{i}+2\sum_{p=1}^{k}\left(\sum_{i=0}^{k-p}\sum_{j=k-p+1}^{k}\beta_{ij}\right)\in(0,1), it follows that A+2C(0,1)A+2C\in(0,1), so the associated operator FF is a continuous convex contraction of order 2 from XX to XX, because

D=d(F2x,F2y)(i=0k1αi)d(x,y)+2p=1k(i=0kpj=kp+1kβij)d(Fx,Fy),D=d\left(F^{2}x,F^{2}y\right)\leq\left(\sum_{i=0}^{k-1}\alpha_{i}\right)d(x,y)+2\cdot\sum_{p=1}^{k}\left(\sum_{i=0}^{k-p}\sum_{j=k-p+1}^{k}\beta_{ij}\right)d(Fx,Fy),

for each arbitrary elements x,yXx,y\in X. So FF has a unique fixed point x=F(x)x^{*}=F\left(x^{*}\right). It follows that x=f(x,,x)x^{*}=f\left(x^{*},\ldots,x^{*}\right), i.e., xx^{*} is the unique fixed point of f:XkXf:X^{k}\rightarrow X.
Finally, from the fact that the Picard sequence of F, i.e., (xn)n\left(x_{n}\right)_{n\in\mathbb{N}}, defined by xn+1=F(xn)x_{n+1}=F\left(x_{n}\right) is convergent to x=F(x)x^{*}=F\left(x^{*}\right), it follows that the sequence defined as xn+1=f(xn,,xn)x_{n+1}=f\left(x_{n},\ldots,x_{n}\right) is convergent to the unique fixed point of f , i.e., x=f(x,,x)x^{*}=f\left(x^{*},\ldots,x^{*}\right), so the proof is over.
(ii) Let f:XkXf:X^{k}\rightarrow X be a continuous Presić convex contraction of the second kind.
As before, we consider the associated operator FF of ff, such as F:XXF:X\rightarrow X and show that is a convex contraction of the second order.
Let x,yXx,y\in X arbitrary elements from the metric space XX. We compute the following distance, applying triangle inequality k times, as in (i).

D\displaystyle D :=d(F2x,F2y)\displaystyle=d\left(F^{2}x,F^{2}y\right)
=d(f(f(x,,x),,f(x,,x)),f(f(y,,y),,f(y,,y)))\displaystyle=d(f(f(x,\ldots,x),\ldots,f(x,\ldots,x)),f(f(y,\ldots,y),\ldots,f(y,\ldots,y)))
D1+D2++Dk1+Dk\displaystyle\leq D_{1}+D_{2}+\cdots+D_{k-1}+D_{k}

We make the remark that we shall apply the triangle inequality in a different way in contrast to the proof of (i). Moreover, we shall use the same terminology as in the proof of (i).
We ought to evaluate each of Di,i=1,k¯D_{i},i=\overline{1,k}, i.e.,

D1=d(f(f(x,,x),,f(x,,x)),\displaystyle D_{1}=d(f(f(x,\ldots,x),\ldots,f(x,\ldots,x)),
f(f(x,,x,y),,f(x,,x,y),f(x,,x,y))),\displaystyle f(f(x,\ldots,x,y),\ldots,f(x,\ldots,x,y),f(x,\ldots,x,y))),

where x0=x1==xk1=xx_{0}=x_{1}=\cdots=x_{k-1}=x and xk=yx_{k}=y.
A similar approach is made for D2D_{2}; also, Dk1D_{k-1} can be computed in the same way.

Dk1=d(f(f(x,x,y,,y),,f(x,x,y,,y)),\displaystyle D_{k-1}=d(f(f(x,x,y,\ldots,y),\ldots,f(x,x,y,\ldots,y)),
f(f(x,y,,y),,f(x,y,,y))),\displaystyle\quad f(f(x,y,\ldots,y),\ldots,f(x,y,\ldots,y))),

where the elements from the convex contractive condition are: x0=x1=xx_{0}=x_{1}=x and x2==xk1=xk=yx_{2}=\cdots=x_{k-1}=x_{k}=y.
Finally, the last element obtained from the triangle inequality is DkD_{k}.
Dk=d(f(f(x,y,,y),f(x,y,,y),,f(x,y,,y))D_{k}=d(f(f(x,y,\ldots,y),f(x,y,\ldots,y),\ldots,f(x,y,\ldots,y)),
f(f(y,y,,y),,f(y,y,,y),f(y,y,,y)))f(f(y,y,\ldots,y),\ldots,f(y,y,\ldots,y),f(y,y,\ldots,y))), with x0=xx_{0}=x and x1=x2==xk1=xk=yx_{1}=x_{2}=\cdots=x_{k-1}=x_{k}=y.

Even though the distances D1,,DkD_{1},\ldots,D_{k} differ from the distances (also obtained from triangle inequality) from (i) of the same theorem, the contractive convex condition of the second kind will finally lead to the same values of D1,,DkD_{1},\ldots,D_{k}, so the condition over the coefficients αi\alpha_{i}, with i=0,k1¯i=\overline{0,k-1} and βij\beta_{ij}, with i,j=0,k¯i,j=\overline{0,k} is the same as in (i). Now, for the second part of the proof, the same observation can be made as in the proof of (i) regarding the Picard iterative sequence. In this way, it follows that the conclusion holds.

Remark 2.5. (a) In the previous proof, we have imposed the condition of symmetry, i.e., βij=βji\beta_{ij}=\beta_{ji}, for each i,j=0,,k¯i,j=\overline{0,\ldots,k}. So, in the examples of the last section, we can take δij\delta_{ij} to be the coefficients of d(Fxi,Fxj)d\left(Fx_{i},Fx_{j}\right). This means that we can put βij=βji=δij2\beta_{ij}=\beta_{ji}=\frac{\delta_{ij}}{2}. This does not restrict the condition of the previous theorem, because the sum of βij\beta_{ij} and βji\beta_{ji} appears in the sum of the previous proof.
(b) In the previous proof, with the above notation, we had that
D=d(F2x,F2y)(i=0k1αi)d(x,y)+p=1k(i=0kpj=kp+1kδij)d(Fx,Fy)D=d\left(F^{2}x,F^{2}y\right)\leq\left(\sum_{i=0}^{k-1}\alpha_{i}\right)d(x,y)+\sum_{p=1}^{k}\left(\sum_{i=0}^{k-p}\sum_{j=k-p+1}^{k}\delta_{ij}\right)d(Fx,Fy).
(c) We can easily put i,j=0,ijk\sum_{i,j=0,i\neq j}^{k} in the Definitions 1.3 and 1.4, since, for i=ji=j the distance d(Fxi,Fxj)=0d\left(Fx_{i},Fx_{j}\right)=0. So it does not matter what values take the diagonal coefficients βii,i=0,k¯\beta_{ii},i=\overline{0,k}.

Now, we give an alternative approach for computing the condition on the coefficients αi\alpha_{i} and βij\beta_{ij} from the previous theorem.

Remark 2.6. The sum CC from the previous theorem can be written as C=Sk+Sk1++S2+S1C=S_{k}+S_{k-1}+\cdots+S_{2}+S_{1}.

C\displaystyle C =i=0k1j=kkβij+i=0k2j=k1kβij++i=01j=2kβij+i=00j=1kβij, where\displaystyle=\sum_{i=0}^{k-1}\sum_{j=k}^{k}\beta_{ij}+\sum_{i=0}^{k-2}\sum_{j=k-1}^{k}\beta_{ij}+\cdots+\sum_{i=0}^{1}\sum_{j=2}^{k}\beta_{ij}+\sum_{i=0}^{0}\sum_{j=1}^{k}\beta_{ij},\text{ where }
S1\displaystyle S_{1} =i=00j=1kβij=j=1kβ0j.\displaystyle=\sum_{i=0}^{0}\sum_{j=1}^{k}\beta_{ij}=\sum_{j=1}^{k}\beta_{0j}.
S2\displaystyle S_{2} =i=01j=2kβij=i=00j=2kβij+j=2kβ1j=j=2ki=00βij+j=2kβ1j\displaystyle=\sum_{i=0}^{1}\sum_{j=2}^{k}\beta_{ij}=\sum_{i=0}^{0}\sum_{j=2}^{k}\beta_{ij}+\sum_{j=2}^{k}\beta_{1j}=\sum_{j=2}^{k}\sum_{i=0}^{0}\beta_{ij}+\sum_{j=2}^{k}\beta_{1j}
=j=1ki=00βij+[j=2kβ1ji=00βi1]=S1+a1.\displaystyle=\sum_{j=1}^{k}\sum_{i=0}^{0}\beta_{ij}+\left[\sum_{j=2}^{k}\beta_{1j}-\sum_{i=0}^{0}\beta_{i1}\right]=S_{1}+a_{1}.
S3\displaystyle S_{3} =i=02j=3kβij=i=01j=3kβij+j=3kβ2j=j=3ki=01βij+j=3kβ2j\displaystyle=\sum_{i=0}^{2}\sum_{j=3}^{k}\beta_{ij}=\sum_{i=0}^{1}\sum_{j=3}^{k}\beta_{ij}+\sum_{j=3}^{k}\beta_{2j}=\sum_{j=3}^{k}\sum_{i=0}^{1}\beta_{ij}+\sum_{j=3}^{k}\beta_{2j}
=j=2ki=01βij+[j=3kβ2ji=01βi2]=S2+a2.\displaystyle=\sum_{j=2}^{k}\sum_{i=0}^{1}\beta_{ij}+\left[\sum_{j=3}^{k}\beta_{2j}-\sum_{i=0}^{1}\beta_{i2}\right]=S_{2}+a_{2}.
Sk1\displaystyle S_{k-1} =i=0k2j=k1kβij=i=0k3j=k1kβij+j=k1k\displaystyle=\sum_{i=0}^{k-2}\sum_{j=k-1}^{k}\beta_{ij}=\sum_{i=0}^{k-3}\sum_{j=k-1}^{k}\beta_{ij}+\sum_{j=k-1}^{k}
βk2,j\displaystyle\beta_{k-2,j} =j=k1ki=0k3βij+j=k1kβk2,j\displaystyle=\sum_{j=k-1}^{k}\sum_{i=0}^{k-3}\beta_{ij}+\sum_{j=k-1}^{k}\beta_{k-2,j}
=j=k2ki=0k3βij+[j=k1kβk2,ji=0k3βi,k2]=Sk2+ak2\displaystyle=\sum_{j=k-2}^{k}\sum_{i=0}^{k-3}\beta_{ij}+\left[\sum_{j=k-1}^{k}\beta_{k-2,j}-\sum_{i=0}^{k-3}\beta_{i,k-2}\right]=S_{k-2}+a_{k-2}
Sk\displaystyle S_{k} =i=0k1j=kkβij=i=0k2j=kkβij+j=kkβk1,j=j=kki=0k2βij+j=kkβk1,j\displaystyle=\sum_{i=0}^{k-1}\sum_{j=k}^{k}\beta_{ij}=\sum_{i=0}^{k-2}\sum_{j=k}^{k}\beta_{ij}+\sum_{j=k}^{k}\beta_{k-1,j}=\sum_{j=k}^{k}\sum_{i=0}^{k-2}\beta_{ij}+\sum_{j=k}^{k}\beta_{k-1,j}
=j=k1ki=0k2βij+[j=kkβk1,ji=0k2βi,k1]=Sk1+ak1\displaystyle=\sum_{j=k-1}^{k}\sum_{i=0}^{k-2}\beta_{ij}+\left[\sum_{j=k}^{k}\beta_{k-1,j}-\sum_{i=0}^{k-2}\beta_{i,k-1}\right]=S_{k-1}+a_{k-1}

So, we get that {S2=S1+a1,S3=S2+a2,Sk1=Sk2+ak2,Sk=Sk1+ak1.\left\{\begin{array}[]{l}S_{2}=S_{1}+a_{1},\\ S_{3}=S_{2}+a_{2},\\ \cdots\\ S_{k-1}=S_{k-2}+a_{k-2},\\ S_{k}=S_{k-1}+a_{k-1}.\end{array}\right.
So Sk=S1+a1++ak2+ak1=j=1kβ0j+a1++ak2+ak1S_{k}=S_{1}+a_{1}+\cdots+a_{k-2}+a_{k-1}=\sum_{j=1}^{k}\beta_{0j}+a_{1}+\cdots+a_{k-2}+a_{k-1}.
Now, it follows that

{Sk=ak1+ak2++a1+S1Sk1=0+ak2++a1+S1S2=0+0++a1+S1S1=0+0++0+S1\left\{\begin{array}[]{l}S_{k}=a_{k-1}+a_{k-2}+\cdots+a_{1}+S_{1}\\ S_{k-1}=0+a_{k-2}+\cdots+a_{1}+S_{1}\\ \cdots\\ S_{2}=0+0+\cdots+a_{1}+S_{1}\\ S_{1}=0+0+\cdots+0+S_{1}\end{array}\right.

Now, the sum CC from the previous theorem had the value C=S1++Sk1+SkC=S_{1}+\cdots+S_{k-1}+S_{k}. So, it follows that C=kS1+(k1)a1+(k2)a2++2ak2+ak1C=kS_{1}+(k-1)a_{1}+(k-2)a_{2}+\cdots+2a_{k-2}+a_{k-1}. Let us denote by a0a_{0} the element S1S_{1}. Then, we get that C=(k0)a0+(k1)a1+(k2)a2++2ak2+ak1C=(k-0)a_{0}+(k-1)a_{1}+(k-2)a_{2}+\cdots+2a_{k-2}+a_{k-1}, where

{ap=j=p+1kβpji=0p1βip, where p=1,k1¯a0=i=00j=1kβij=j=1kβ0j\left\{\begin{array}[]{l}a_{p}=\sum_{j=p+1}^{k}\beta_{pj}-\sum_{i=0}^{p-1}\beta_{ip},\text{ where }p=\overline{1,k-1}\\ a_{0}=\sum_{i=0}^{0}\sum_{j=1}^{k}\beta_{ij}=\sum_{j=1}^{k}\beta_{0j}\end{array}\right.

With the convention that 01()=0\sum_{0}^{-1}(\ldots)=0, the coefficient a0a_{0} can be written as a0=j=1kβ0ji=01βi0a_{0}=\sum_{j=1}^{k}\beta_{0j}-\sum_{i=0}^{-1}\beta_{i0}, so the formula (2.2) remains valid for p=0,k1¯p=\overline{0,k-1}.

So, for p=0,k1¯p=\overline{0,k-1}, it follows that C=p=0k1(kp)apC=\sum_{p=0}^{k-1}(k-p)a_{p}. Then

C=p=0k1[(j=p+1kβpji=0p1βip)(kp)].C=\sum_{p=0}^{k-1}\left[\left(\sum_{j=p+1}^{k}\beta_{pj}-\sum_{i=0}^{p-1}\beta_{ip}\right)\cdot(k-p)\right].

Finally, the contractive-type condition from (i) of Theorem 2.4 is

(i=0k1αi)+2p=0k1[(j=p+1kβpji=0p1βip)(kp)][0,1).\left(\sum_{i=0}^{k-1}\alpha_{i}\right)+2\sum_{p=0}^{k-1}\left[\left(\sum_{j=p+1}^{k}\beta_{pj}-\sum_{i=0}^{p-1}\beta_{ip}\right)\cdot(k-p)\right]\in[0,1).

Corollary 2.7. If we take k=1k=1, Presić convex contractions of the first kind and Presić convex contractions of the second kind satisfy

d(f2x0,f2x1)α0d(x0,x1)+[β01+β10]d(fx0,fx1),d\left(f^{2}x_{0},f^{2}x_{1}\right)\leq\alpha_{0}d\left(x_{0},x_{1}\right)+\left[\beta_{01}+\beta_{10}\right]d\left(fx_{0},fx_{1}\right),

with α0+[β10+β01][0,1)\alpha_{0}+\left[\beta_{10}+\beta_{01}\right]\in[0,1) for each x0x_{0} and x1x_{1} arbitrary elements of XX. This means that our new types of Presić mappings are proper generalizations of convex contractions of second order.

In the next remark, we show that Presić operators are Presić convex contractions of the first kind. This remark leads to the fact that Presić convex contractions of the first kind are true generalizations of Presić mappings.

Remark 2.8. Let us define a Presić contraction on XX, where ( X,dX,d ) is a metric space and f:XkXf:X^{k}\rightarrow X a mapping satisfying

d(f(w0,,wk1),f(w1,,wk))α0d(w0,w1)++α1d(wk1,wk),d\left(f\left(w_{0},\ldots,w_{k-1}\right),f\left(w_{1},\ldots,w_{k}\right)\right)\leq\alpha_{0}d\left(w_{0},w_{1}\right)+\cdots+\alpha_{1}d\left(w_{k-1},w_{k}\right),

where w0,,wkw_{0},\ldots,w_{k} are arbitrary elements of XX.
Let x0,,xkx_{0},\ldots,x_{k} be also arbitrary elements of the metric space ( X,dX,d ).
Let us choose w0,w1,,wk1w_{0},w_{1},\ldots,w_{k-1} and wkw_{k} in the following way :

{w0=f(x0,,x0),wk=f(xk,,xk),\left\{\begin{array}[]{l}w_{0}=f\left(x_{0},\ldots,x_{0}\right),\\ \ldots\\ w_{k}=f\left(x_{k},\ldots,x_{k}\right),\end{array}\right.

where x0,,xkx_{0},\ldots,x_{k} are arbitrary elements of XX. Then, the Presić mapping on these elements satisfies

Δ:=\displaystyle\Delta= d(f(f(x0,,x0),,f(xk1,,xk1)),\displaystyle d\left(f\left(f\left(x_{0},\ldots,x_{0}\right),\ldots,f\left(x_{k-1},\ldots,x_{k-1}\right)\right),\right.
f(f(x1,,x1),,f(xk,,xk)))\displaystyle\left.f\left(f\left(x_{1},\ldots,x_{1}\right),\ldots,f\left(x_{k},\ldots,x_{k}\right)\right)\right)
\displaystyle\leq i=0k1αid(f(xi,,xi),f(xi+1,,xi+1))=\displaystyle\sum_{i=0}^{k-1}\alpha_{i}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{i+1},\ldots,x_{i+1}\right)\right)=
=\displaystyle= i=0k1αi2d(f(xi,,xi),f(xi+1,,xi+1))+\displaystyle\sum_{i=0}^{k-1}\frac{\alpha_{i}}{2}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{i+1},\ldots,x_{i+1}\right)\right)+
+i=0k1αi2d(f(xi,,xi),f(xi+1,,xi+1)).\displaystyle+\sum_{i=0}^{k-1}\frac{\alpha_{i}}{2}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{i+1},\ldots,x_{i+1}\right)\right).

Since ff is a Presić mapping, it follows that

d(f(xi,,xi),f(xi+1,,xi+1))D1+D2++Dk1+Dk,d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{i+1},\ldots,x_{i+1}\right)\right)\leq D_{1}+D_{2}+\cdots+D_{k-1}+D_{k},
{D1=d(f(xi,,xi),f(xi,,xi,xi+1))D2=d(f(xi,,xi,xi+1),f(xi,,xi,xi+1,xi+1))Dk=d(f(xi,xi+1,,xi+1),f(xi+1,,xi+1,xi+1))\left\{\begin{array}[]{l}D_{1}=d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{i},\ldots,x_{i},x_{i+1}\right)\right)\\ D_{2}=d\left(f\left(x_{i},\ldots,x_{i},x_{i+1}\right),f\left(x_{i},\ldots,x_{i},x_{i+1},x_{i+1}\right)\right)\\ \cdots\\ D_{k}=d\left(f\left(x_{i},x_{i+1},\ldots,x_{i+1}\right),f\left(x_{i+1},\ldots,x_{i+1},x_{i+1}\right)\right)\end{array}\right.

Now, the Presić condition is applied when the coefficients w0,,wkw_{0},\ldots,w_{k} are chosen in the following way : in the case of the distance D1D_{1}, we take w0==wk1=xiw_{0}=\cdots=w_{k-1}=x_{i} and wk=xi+1w_{k}=x_{i+1}, and in the case of the distance D2D_{2}, we take w0==wk2=xiw_{0}=\cdots=w_{k-2}=x_{i} and wk1=wk=xi+1w_{k-1}=w_{k}=x_{i+1}. Also, for the last case, that of DkD_{k}, we take w0=xiw_{0}=x_{i} and w1==wk=xi+1w_{1}=\cdots=w_{k}=x_{i+1}.
So, D1,,DkD_{1},\ldots,D_{k} satisfy

D1αk1d(xi,xi+1),D2αk2d(xi,xi+1),,Dkα0d(xi,xi+1)D_{1}\leq\alpha_{k-1}d\left(x_{i},x_{i+1}\right),D_{2}\leq\alpha_{k-2}d\left(x_{i},x_{i+1}\right),\ldots,D_{k}\leq\alpha_{0}d\left(x_{i},x_{i+1}\right)

Let us denote by δij=βij2\delta_{ij}=\frac{\beta_{ij}}{2}. We have used the fact that, in Theorem 2.4, the symmetry implied βij=βji\beta_{ij}=\beta_{ji}.
Now, from the proof of Theorem 2.4, we know that

C\displaystyle C =i=0k1j=kkβij+i=0k2j=k1kβij++i=01j=2kβij+i=00j=1kβij\displaystyle=\sum_{i=0}^{k-1}\sum_{j=k}^{k}\beta_{ij}+\sum_{i=0}^{k-2}\sum_{j=k-1}^{k}\beta_{ij}+\cdots+\sum_{i=0}^{1}\sum_{j=2}^{k}\beta_{ij}+\sum_{i=0}^{0}\sum_{j=1}^{k}\beta_{ij}
=Sk++S1\displaystyle=S_{k}+\cdots+S_{1}

where Sk=i=0k1j=kkβijS_{k}=\sum_{i=0}^{k-1}\sum_{j=k}^{k}\beta_{ij}. Since we are interested in the coefficients βi,i+1\beta_{i,i+1}, taking j=i+1j=i+1, we obtain that i=0k1k1k1βi,i+1\sum_{i=0}^{k-1}\sum_{k-1}^{k-1}\beta_{i,i+1} leads to Sk=βk1,kS_{k}=\beta_{k-1,k}, because in this case we have that i=k1i=k-1 (this is the common term of the sums).
Also, we know that S1=i=00j=1kβijS_{1}=\sum_{i=0}^{0}\sum_{j=1}^{k}\beta_{ij}, so since we are interested in the term βi,i+1\beta_{i,i+1}, taking j=i+1j=i+1 in S1S_{1}, it follows that i=00i=0kβi,i+1\sum_{i=0}^{0}\sum_{i=0}^{k}\beta_{i,i+1} leads to the common term of the sums, so i=0i=0. This means that S1=β0,1S_{1}=\beta_{0,1}. Now, we denote by S=2CS=2C. From the condition of Theorem 2.4, it follows that A+2C<1A+2C<1, i.e., A+S<1A+S<1, where A=i=0k1αi2A=\sum_{i=0}^{k-1}\frac{\alpha_{i}}{2} and S=2[β01++βk1,k]=[δ01++δk1,k]S=2\left[\beta_{01}+\cdots+\beta_{k-1,k}\right]=\left[\delta_{01}+\cdots+\delta_{k-1,k}\right].
It follows that

Δ\displaystyle\Delta\leq i=0k1αi2(D1++Dk)+i=0k1αi2d(f(xi,,xi),f(xi+1,,xi+1))\displaystyle\sum_{i=0}^{k-1}\frac{\alpha_{i}}{2}\left(D_{1}+\cdots+D_{k}\right)+\sum_{i=0}^{k-1}\frac{\alpha_{i}}{2}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{i+1},\ldots,x_{i+1}\right)\right)
Δ\displaystyle\Delta\leq i=0k1αi[αk1++α0]d(xi,xi+1)\displaystyle\sum_{i=0}^{k-1}\alpha_{i}\left[\alpha_{k-1}+\cdots+\alpha_{0}\right]d\left(x_{i},x_{i+1}\right)
+i=0k1αi2d(f(xi,,xi),f(xi+1,,xi+1))\displaystyle+\sum_{i=0}^{k-1}\frac{\alpha_{i}}{2}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{i+1},\ldots,x_{i+1}\right)\right)
=\displaystyle= i=0k1αi2t=0k1αt<1d(xi,xi+1)+i=0k1αi2d(f(xi,,xi),f(xi+1,,xi+1))\displaystyle\sum_{i=0}^{k-1}\frac{\alpha_{i}}{2}\cdot\underbrace{\sum_{t=0}^{k-1}\alpha_{t}}_{<1}\cdot d\left(x_{i},x_{i+1}\right)+\sum_{i=0}^{k-1}\frac{\alpha_{i}}{2}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{i+1},\ldots,x_{i+1}\right)\right)
<i=0k1αi2d(xi,xi+1)+i=0k1αi2d(f(xi,,xi),f(xi+1,,xi+1)).<\sum_{i=0}^{k-1}\frac{\alpha_{i}}{2}d\left(x_{i},x_{i+1}\right)+\sum_{i=0}^{k-1}\frac{\alpha_{i}}{2}d\left(f\left(x_{i},\ldots,x_{i}\right),f\left(x_{i+1},\ldots,x_{i+1}\right)\right).

We infer that taking δi,i+1=αi2\delta_{i,i+1}=\frac{\alpha_{i}}{2} and δij=0\delta_{ij}=0 for i,j\mathrm{i},\mathrm{j} nonconsecutive index, for each i,j=0,k¯i,j=\overline{0,k} in Theorem 2.4, the Presić mapping is indeed a convex contraction of the first kind, since αi<1\alpha_{i}<1, for each i{0,,k1}i\in\{0,\ldots,k-1\}.

Now, we present a data dependence theorem for Presić-type convex contractions. We shall employ additional condition for the function ff as in [12]. This theorem concerns data dependence results related to Presić convex contractions of the first kind, respectively Presić convex contractions of the second kind.

Theorem 2.9. Let ( X,dX,d ) be a complete metric space. Also, let g:XkXg:X^{k}\rightarrow X be a mapping with at least a fixed point xgXx_{g}^{*}\in X.
(i) Let f:XkXf:X^{k}\rightarrow X be a continuous Presić convex contraction of the first kind, satisfying the conditions from (i) of Theorem 2.4.
Furthermore, suppose the following assumptions are also satisfied :
(a) there exists η1>0\eta_{1}>0, such that d(f(x,,x),g(x,,x))η1d(f(x,\ldots,x),g(x,\ldots,x))\leq\eta_{1}, for each xXx\in X,
(b) there exists η2>0\eta_{2}>0, such that for each xXx\in X, we have that

d(f(f(x,,x),,f(x,,x)),g(g(x,,x),,g(x,,x)))η2Thend(xf,xg)η2+2p=1ki=0kp(j=kp+1kβij)η11[(i=0k1αi)+2p=1ki=0kp(j=kp+1kβij)],\begin{aligned} &d(f(f(x,\ldots,x),\ldots,f(x,\ldots,x)),g(g(x,\ldots,x),\ldots,g(x,\ldots,x)))\leq\eta_{2}\\ &\operatorname{Thend}\left(x_{f}^{*},x_{g}^{*}\right)\leq\frac{\eta_{2}+2\sum_{p=1}^{k}\sum_{i=0}^{k-p}\left(\sum_{j=k-p+1}^{k}\beta_{ij}\right)\cdot\eta_{1}}{1-\left[\left(\sum_{i=0}^{k-1}\alpha_{i}\right)+2\sum_{p=1}^{k}\sum_{i=0}^{k-p}\left(\sum_{j=k-p+1}^{k}\beta_{ij}\right)\right]}\end{aligned},

where xfx_{f}^{*} is the unique fixed point of ff.
(ii) Consider f:XkXf:X^{k}\rightarrow X a continuous Presić convex contraction of the second kind, satisfying the conditions from (ii) of Theorem 2.4.
In a similar way, suppose that ff and gg satisfy the previous conditions (a) and (b). Then, d(xf,xg)d\left(x_{f}^{*},x_{g}^{*}\right) has the same major bound as before.

Proof. (i) Using triangle inequality applied kk times, we get that

d(xf,xg)=d(f(xf,,xf),g(xg,,xg))\displaystyle d\left(x_{f}^{*},x_{g}^{*}\right)=d\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),g\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)
=d(f(f(xf,,xf),,f(xf,,xf)),g(g(xg,,xg),,g(xg,,xg)))\displaystyle\quad=d\left(f\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),\ldots,f\left(x_{f}^{*},\ldots,x_{f}^{*}\right)\right),g\left(g\left(x_{g}^{*},\ldots,x_{g}^{*}\right),\ldots,g\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)\right)
d(f(f(xf,,xf),,f(xf,,xf)),f(f(xg,,xg),,f(xg,,xg)))\displaystyle\quad\leq d\left(f\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),\ldots,f\left(x_{f}^{*},\ldots,x_{f}^{*}\right)\right),f\left(f\left(x_{g}^{*},\ldots,x_{g}^{*}\right),\ldots,f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)\right)
+d(f(f(xg,,xg),,f(xg,,xg)),g(g(xg,,xg),,g(xg,,xg)))\displaystyle\quad+d\left(f\left(f\left(x_{g}^{*},\ldots,x_{g}^{*}\right),\ldots,f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right),g\left(g\left(x_{g}^{*},\ldots,x_{g}^{*}\right),\ldots,g\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)\right)
η2+d(f(f(xf,,xf),,f(xf,,xf)),f(f(xg,,xg),,f(xg,,xg)))\displaystyle\quad\leq\eta_{2}+d\left(f\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),\ldots,f\left(x_{f}^{*},\ldots,x_{f}^{*}\right)\right),f\left(f\left(x_{g}^{*},\ldots,x_{g}^{*}\right),\ldots,f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)\right)
η2+D, where DD1++Dk\displaystyle\quad\leq\eta_{2}+D,\text{ where }D\leq D_{1}+\cdots+D_{k}
 with D=d(f(f(xf,,xf),,f(xf,,xf)),f(f(xg,,xg),,\displaystyle\quad\text{ with }D=d\left(f\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),\ldots,f\left(x_{f}^{*},\ldots,x_{f}^{*}\right)\right),f\left(f\left(x_{g}^{*},\ldots,x_{g}^{*}\right),\ldots,\right.\right.
f(xg,,xg)))\displaystyle\left.\left.\quad f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)\right)

Now, the first distance D1D_{1} is

D1=d(f(f(xf,,xf),,f(xf,,xf))\displaystyle D_{1}=d\left(f\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),\ldots,f\left(x_{f}^{*},\ldots,x_{f}^{*}\right)\right)\right.
f(f(xf,,xf),,f(xf,,xf),f(xg,,xg)))\displaystyle\left.f\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),\ldots,f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)\right)

The second distance D2D_{2} is

D2=d(f(f(xf,,xf),,f(xf,,xf),f(xg,,xg))\displaystyle D_{2}=d\left(f\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),\ldots,f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)\right.
f(f(xf,,xf),,f(xf,,xf),f(xg,,xg),f(xg,,xg)))\displaystyle\left.\quad f\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),\ldots,f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),f\left(x_{g}^{*},\ldots,x_{g}^{*}\right),f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)\right)

Finally, the last distance DkD_{k} is

Dk=d(f(f(xf,,xf),f(xg,,xg),,f(xg,,xg),f(xg,,xg)),\displaystyle D_{k}=d\left(f\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),f\left(x_{g}^{*},\ldots,x_{g}^{*}\right),\ldots,f\left(x_{g}^{*},\ldots,x_{g}^{*}\right),f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right),\right.
f(f(xg,,xg),,f(xg,,xg),f(xg,,xg))).\displaystyle\left.f\left(f\left(x_{g}^{*},\ldots,x_{g}^{*}\right),\ldots,f\left(x_{g}^{*},\ldots,x_{g}^{*}\right),f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)\right).

Now, applying the convex-contractive condition, with

{x0==xk1=xf and xk=xg for D1x0==xk2=xf and xk1=xk=xg for D2x0=xf and x1==xk=xg for Dk\left\{\begin{array}[]{l}x_{0}=\ldots=x_{k-1}=x_{f}^{*}\text{ and }x_{k}=x_{g}^{*}\text{ for }D_{1}\\ x_{0}=\ldots=x_{k-2}=x_{f}^{*}\text{ and }x_{k-1}=x_{k}=x_{g}^{*}\text{ for }D_{2}\\ \cdots\\ x_{0}=x_{f}^{*}\text{ and }x_{1}=\ldots=x_{k}=x_{g}^{*}\text{ for }D_{k}\end{array}\right.

we obtain, by the same reasoning as in Theorem 2.4, with x=xfx=x_{f}^{*} and y=xgy=x_{g}^{*}, that

D\displaystyle D (i=0k1αi)d(xf,xg)+2p=1k(i=0kpj=kp+1kβij)D\displaystyle\leq\left(\sum_{i=0}^{k-1}\alpha_{i}\right)d\left(x_{f}^{*},x_{g}^{*}\right)+2\sum_{p=1}^{k}\left(\sum_{i=0}^{k-p}\sum_{j=k-p+1}^{k}\beta_{ij}\right)D^{\prime}
=Ad(xf,xg)+2CD\displaystyle=Ad\left(x_{f}^{*},x_{g}^{*}\right)+2C\cdot D^{\prime}

with the notation from the proof of Theorem 2.4,

where D=d(f(xf,,xf),f(xg,,xg))\displaystyle\text{ where }D^{\prime}=d\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)
d(f(xf,,xf),g(xg,,xg))+d(g(xg,,xg),f(xg,,xg))\displaystyle\leq d\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),g\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)+d\left(g\left(x_{g}^{*},\ldots,x_{g}^{*}\right),f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)
η1+d(xf,xg)\displaystyle\leq\eta_{1}+d\left(x_{f}^{*},x_{g}^{*}\right)
So d(xf,xg)η2+(i=0k1αi)d(xf,xg)+2p=1k(i=0kpj=kp+1kβij)η1\displaystyle\text{ So }d\left(x_{f}^{*},x_{g}^{*}\right)\leq\eta_{2}+\left(\sum_{i=0}^{k-1}\alpha_{i}\right)d\left(x_{f}^{*},x_{g}^{*}\right)+2\sum_{p=1}^{k}\left(\sum_{i=0}^{k-p}\sum_{j=k-p+1}^{k}\beta_{ij}\right)\cdot\eta_{1}
+2p=1k(i=0kpj=kp+1kβij)d(xf,xg).\displaystyle+2\sum_{p=1}^{k}\left(\sum_{i=0}^{k-p}\sum_{j=k-p+1}^{k}\beta_{ij}\right)\cdot d\left(x_{f}^{*},x_{g}^{*}\right)\text{. }

Finally, taking d(xf,xg)d\left(x_{f}^{*},x_{g}^{*}\right) in the left hand side and using the fact that

i=0k1αi+2p=1k(i=0kpj=kp+1kβij)<1, we get the conclusion. \sum_{i=0}^{k-1}\alpha_{i}+2\sum_{p=1}^{k}\left(\sum_{i=0}^{k-p}\sum_{j=k-p+1}^{k}\beta_{ij}\right)<1,\text{ we get the conclusion. }

(ii) Applying the triangle inequality and using the fact that xfx_{f}^{*} is the unique fixed point of ff and that xgx_{g}^{*} is a fixed point for the mapping gg, we obtain, as in the previous proof, that

D=d(f(f(xf,,xf),,f(xf,,xf)),f(f(xg,,xg),,D=d\left(f\left(f\left(x_{f}^{*},\ldots,x_{f}^{*}\right),\ldots,f\left(x_{f}^{*},\ldots,x_{f}^{*}\right)\right),f\left(f\left(x_{g}^{*},\ldots,x_{g}^{*}\right),\ldots,\right.\right.
f(xg,,xg))).\left.\left.f\left(x_{g}^{*},\ldots,x_{g}^{*}\right)\right)\right).

Now, as in the proof of (ii) of Theorem 2.4, applying the convex contractivetype condition on ff, we get the same as conclusion as in the first data dependence result.

For the case when k=1k=1, the data dependence result becomes the following.
Corollary 2.10. Taking k=1k=1 in the previous theorem, we obtain that 2p=1k(i=0kpj=kp+1kβij)=2β1,0=β1,0+β0,12\sum_{p=1}^{k}\left(\sum_{i=0}^{k-p}\sum_{j=k-p+1}^{k}\beta_{ij}\right)=2\beta_{1,0}=\beta_{1,0}+\beta_{0,1}. It follows that the data dependence results are valid for the case of k=1k=1 as in [12], where d(xf,xg)η2+δη11(α+δ)d\left(x_{f}^{*},x_{g}^{*}\right)\leq\frac{\eta_{2}+\delta\cdot\eta_{1}}{1-(\alpha+\delta)}, where δ:=2β10\delta:=2\beta_{10} and α:=α0\alpha:=\alpha_{0}.

3. Examples

Concerning Presić-type operators, very few non-trivial examples can be found in the literature and most of these examples are piecewise, constant or linear mappings from X2X^{2} to XX, where ( X,dX,d ) is a metric space.
In this section, we shall present non-trivial examples of continuous mappings that satisfy the conditions from (Theorem 2.4). The self-mappings under consideration are defined on an interval [0,r][0,r], with r>0r>0 and x=0x^{*}=0 is their unique fixed point. Furthermore, we present situations such that this particular mappings are not Presić operators. Also, other situations in which we modify the coefficients so our operators can be Presić mappings are given. Our first two results of the present section are examples of a mapping satisfying the conditions from (i) of Theorem 2.4, i.e., the construction of a Presić convex contraction of the first kind that is not a Presić mapping, respectively, and example of Presić convex contraction of the first kind that is a Presić mapping.
Example 3.1. Taking f:X2Xf:X^{2}\rightarrow X, such that f(x,y)=(xy)2af(x,y)=\frac{(xy)^{2}}{a}, where x,y[0,r]x,y\in[0,r], with r=2r=2 and a=12a=12, then ff is an example of a Presić convex contraction of the first kind that is not a Presić mapping.
Example 3.2. Taking f:X2Xf:X^{2}\rightarrow X, such that f(x,y)=(xy)2af(x,y)=\frac{(xy)^{2}}{a}, where x,y[0,r]x,y\in[0,r], with r=2r=2 and a=34a=34, then ff is an example of a Presić convex contraction of the first kind that is a Presić mapping.

Our next two examples concern a mapping that satisfies the conditions from (ii) of Theorem 2.4. Furthermore, the first example regards a mapping that is a convex contraction of the second kind that is not a Presić operator and the second one concerns a mapping that is a convex contraction of the second kind that is at the same time a Presic operator.
Example 3.3. Consider the function f:X2Xf:X^{2}\rightarrow X, defined as f(x,y)=αx2+βy2f(x,y)=\alpha x^{2}+\beta y^{2} and x,yX=[0,r]x,y\in X=[0,r], with r=2,α=0.03r=2,\alpha=0.03 and β=0.27\beta=0.27. Then, ff is a convex contraction of the second kind, but is not a Presić operator.

Example 3.4. Consider the function f:X2Xf:X^{2}\rightarrow X, defined as f(x,y)=αx2+βy2f(x,y)=\alpha x^{2}+\beta y^{2} and x,yX=[0,r]x,y\in X=[0,r], with r=3,α=145r=3,\alpha=\frac{1}{45} and β=17\beta=\frac{1}{7}. Then, ff is a convex contraction of the second kind that is also a Presić operator.

Our last results from this paper are another two examples for convex contraction of Presić-type of the first kind. Moreover, the following example regards a mapping that is a convex contraction of the first kind that is not a Presić operator; meanwhile, the last example concerns a mapping that is a convex contraction of the first kind that is also a Presić-type operator.

Example 3.5. Taking the mapping f:X2Xf:X^{2}\rightarrow X, defined as f(x,y)=(τ1xτ2y)2f(x,y)=\left(\tau_{1}x-\right.\left.\tau_{2}y\right)^{2}, with X=[0,r]X=[0,r], where r=2,τ1=0.6r=2,\tau_{1}=0.6 and τ2=0.8\tau_{2}=0.8, then the mapping ff is not of Presić type, but is a convex contraction of the first kind.

Example 3.6. Taking the mapping f:X2Xf:X^{2}\rightarrow X, defined as f(x,y)=(τ1xτ2y)2f(x,y)=\left(\tau_{1}x-\right.\left.\tau_{2}y\right)^{2}, with X=[0,r]X=[0,r], where r=2,τ1=0.45r=2,\tau_{1}=0.45 and τ2=14+24\tau_{2}=-\frac{1}{4}+\frac{\sqrt{2}}{4}\approx 0.1035533905 , then the mapping ff is a Presić type mapping and also a convex contraction of the first kind.

Acknowledgements

The author is thankful to Dr. Adrian Viorel from Technical University of ClujNapoca for important remarks and ideas that improved the present article. Last, but not least, the author is deeply indebted to the referees for valuable remarks that shaped the structure of this research article.

References

[1] Abbas, M., Ilić, D., Nazir, T.: Iterative approximation of fixed points of generalized weak Presic type kk-step iterative method for a class of operators. Filomat 29(4), 713-724 (2015)
[2] Alghamdi, M.A., Alnafei, S.H., Radenovic, S., Shahzad, N.: Fixed point theorem for convex contraction mappings on cone metric spaces. Math. Comput. Model. 54, 2020-2026 (2011)
[3] Alnafei, S.H., Radenovic, S., Shahzad, N.: Fixed point theorems for mappings with convex diminishing diameters on cone metric spaces. Appl. Math. Lett. 24, 2162-2166 (2011)
[4] Banach, S.: Sur les opérations dans les ensembles abstraits et leur application aux equations integrales. Fund. Math. 3, 133-181 (1922)
[5] Berinde, V., Păcurar, M.: Stability of kk-step fixed point iterative methods for some Prešić type contractive mappings. J. Inequal. Appl. 2014:149 (2014)
[6] Ćirić, L.B., Prešić, L.B.: On Presic type generalisation of Banach contraction mapping principle. Acta Math. Univ. Comenian 76(2), 143-147 (2007)
[7] Istrăţescu, V.: Some fixed point theorems for convex contraction mappings and convex nonexpansive mappings (I). Lib. Math. 1, 151-164 (1981)
[8] Istrăţescu, V.: Some fixed point theorems for convex contraction mappings and mappings with convex diminishing diameters - I. Ann. Mat. Pura Appl. 130, 89-104 (1982)
[9] Istrăţescu, V.: Some fixed point theorems for convex contraction mappings and mappings with convex diminishing diameters. II. Ann. Mat. Pura Appl. 134, 327-362 (1983)
[10] Khan, M.S., Berzig, M., Samet, B.: Some convergence results for iterative sequences of Prešić type and applications. Adv. Differ. Equ. 38 (2012). https:// doi.org/10.1186/1687-1847-2012-38
[11] Miculescu, R., Mihail, A.: A generalization of Matkowski’s fixed point theorem and Istrăţescu’s fixed point theorem concerning convex contractions. J. Fixed Point Theory Appl. 19, 1525-1533 (2017)
[12] Mureşan, V., Mureşan, A.S.: On the theory of fixed point theorems for convex contraction mappings. Carpatian J. Math. 31(3), 365-371 (2015)
[13] Păcurar, M.: Approximating common fixed points of Presić-Kannan type operators by a multi-step iterative method. An. Şt. Univ. Ovidiu Constanţa 17(1), 153-168 (2009)
[14] Păcurar, M.: A multi-step method for approximating common fixed points of Presić-Rus type operators on metric spaces. Studia Univ. Babeş-Bolyai Math. LV(1), 149-162 (2010)
[15] Păcurar, M.: Fixed points of almost Presić operators by a kk-step iterative method. An. Ştiint. Univ. AI.I. Cuza Iaşi Mat. 57:199-210 (2011)
[16] Pathak, H.K., George, R., Nabwey, H.A., El-Paoumy, M.S., Reshma, K.P.: Some generalized fixed point results in bb-metric space and application to matrix equations. Fixed Point Theory Appl. 2015:101 (2015)
[17] Prešić, S.B.: Sur une classe d’inéquations aux differences finies et sur la convergence de certain suites. Pub. de l’Inst. Math. Belgrade 5(19), 75-78 (1965)
[18] Rus, I.: An iterative method for the solution of the equation x=f(x,,x)x=f(x,\ldots,x). Anal. Numér. Thor. Approx. 10(1), 95-100 (1981)
[19] Sastry, K.P.R., Rao, ChS, Sekhar, C., Balaiah, M.: A fixed point theorem for cone convex contractions of order m2m\geq 2. Int. J. Math. Sci. Eng. Appl. 6(1), 263-271 (2012)
[20] Shukla, S., Radenovic, S.: Presić-Maya type theorems in ordered metric spaces. Gulf J. Math. 2(2), 73-82 (2014)
[21] Shukla, S., Radenovic, S., Pantelić, S.: Some fixed point theorems for Prešić-Hardy-Rogers type contractions in metric spaces. J. Math. (2013). https:// doi.org/10.1155/2013/295093
[22] Shukla, S., Radenovic, S.: Some generalizations of Prešić type mappings and applications. An. Ştiinţ. Univ. AI.I. Cuza Iaşi Mat. (2015). https://doi.org/10. 1515/aicu-2015-0026

Cristian Daniel Alecsa
Department of Mathematics
Babeş-Bolyai University
Cluj-Napoca
Romania
and

Tiberiu Popoviciu Institute of Numerical Analysis, Romanian Academy
Cluj-Napoca
Romania
e-mail: cristian.alecsa@math.ubbcluj.ro;
cristian.alecsa@ictp.acad.ro

2018

Related Posts