Journal of Computational and Applied Mathematics 239 (2013) 103–113

Contents lists available at SciVerse ScienceDirect

Journal of Computational and Applied

Mathematics
journal homepage: www.elsevier.com/locate/cam

Common fixed point theorems for P D -operator pairs under relaxed

conditions with applications
H.K. Pathak, Deepmala ∗
School of Studies in Mathematics, Pt. Ravishankar Shukla University, Raipur (C.G.), 492010, India

article info abstract

Article history: In this paper, we define P D -operator pair of single valued mappings and obtain some
Received 5 May 2011 common fixed point theorems for this class of maps under relaxed conditions. Our
Received in revised form 23 March 2012 theorem generalizes results of Bhatt et al. [A. Bhatt, H. Chandra, D.R. Sahu, Common fixed
point theorems for occasionally weakly compatible mappings under relaxed conditions,
MSC: Nonlinear Anal. 73 (2010) 176–182.], Jungck and Rhoades [G. Jungck, B.E. Rhoades, Fixed
54H25
point theorems for occasionally weakly compatible mappings, Fixed Point Theory 7 (2)
47H10
(2006) 287–296.] and Hussain et al. [N. Hussain, M.A. Khamsi, A. Latif, Common fixed
Keywords: points for JH -operators and occasionally weakly biased pairs under relaxed conditions,
Common fixed point Nonlinear Anal. 74 (2011) 2133–2140.] and others. As applications, we also study the
P -operator pair solution of functional equations arising from dynamic programming and variational
JH -operator pair inequalities arises in two point obstacle problem. We also discuss some illustrative
Occasionally weakly compatible maps
examples to highlight the realized improvements.
Dynamic programming
Variational inequality
© 2012 Elsevier B.V. All rights reserved.

1. Introduction

The study of common fixed points of mappings satisfying some contractive type condition have been obtained by various
authors. In 1976, Jungck [1] initiated a study of common fixed points of commuting maps. On the other hand in 1982, Sessa [2]
defined weak commutativity and proved common fixed point theorem for weak commuting mappings. In 1986, Jungck [3]
defined the compatible maps and has been useful as a tool for obtaining the fixed point theorems and in the study of periodic
points. Al-Thagafi and Shahzad [4] defined the concept of occasionally weakly compatible (owc) maps. Recently, Pathak
and Hussain [5], defined the concept of P -operator pair, which contain weakly compatible and owc selfmaps as a proper
class.
In this paper, we introduce the notion of P D -operator pair for single valued maps, which is different from P -operator
pair [5] and JH -operator pair [6]. This contains commuting maps, a nontrivial weakly compatible and nontrivial owc
maps as a proper subclass. For this new operator, we prove some common fixed point theorems for a pair of two single
valued mappings and a quadruple of single valued mappings without using the E.A. property and completeness and weak
compatibility, under a relaxed condition on d. In Section 4, we prove the fixed point theorems without using the concept of
triangle inequality or symmetry and in Section 5, we prove the results in symmetric space only. The results obtained by us
give proper generalization of some important fixed point theorems (see [4–12] and others) and open up a wider scope for
the study of common fixed points under contractive type conditions and discuss some illustrative examples. We also see
the applications of our results in dynamic programming and variational inequality.

∗ Corresponding author. Tel.: +91 771 2262649.

E-mail addresses: hkpathak05@gmail.com (H.K. Pathak), dmrai23@gmail.com (Deepmala).

0377-0427/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.cam.2012.09.008
104 H.K. Pathak, Deepmala / Journal of Computational and Applied Mathematics 239 (2013) 103–113

2. Preliminaries

Throughout this paper, (X , d) denotes a metric space. For x ∈ X and A ⊆ X ,

d(x, A) = inf{d(x, y) : y ∈ A}.

Let f : X −→ X and g : X −→ X . A point x ∈ X is a fixed point of f if x = fx. The set of all fixed points of f is denoted by
F (f ). A point x ∈ X is a coincidence point of f and g if fx = gx. We shall call w = fx = gx a point of coincidence of f and
g. The set of all coincidence points and points of coincidence of f and g are denoted by C (f , g ) and PC (f , g ) respectively. A
point x ∈ X is a common fixed point of f and g if x = fx = gx. The set of all common fixed points of f and g is denoted by
F (f , g ). The pair {f , g } is called

1. commuting [1], if fgx = gfx, ∀x ∈ X .

2. weakly commuting [2], if d(fgx, gfx) ≤ d(fx, gx), for each x ∈ X .
3. compatible [3], if limn d(fgxn , gfxn ) = 0, whenever {xn } is a sequence in X such that limn fxn = limn gxn = t, for some
t ∈ X.
4. weakly compatible [13], if they commute at there coincidence point i.e. fgx = gfx, whenever fx = gx, for x ∈ X .
5. occasionally weakly compatible ([4], see also [8]) (owc), if fgu = gfu, for some u ∈ C (f , g ).
6. P -operator pair [5], if d(u, fu) ≤ diam(C (f , g )), for some u ∈ C (f , g ).
7. JH -operator pair [6], if there is a point u in X such that u ∈ C (f , g ) and d(u, fu) ≤ diam(PC (f , g )), for some u ∈ C (f , g ).

Definition 2.1. Let X be a non-empty set and d be a function d : X × X −→ [0, ∞) such that

d(x, y) = 0 iff x = y, ∀x, y ∈ X . (1)

For a space (X , d) satisfying (1) and A ⊆ X , the diameter of A is defined by

diam(A) = sup{max{d(x, y), d(y, x)} : x, y ∈ A}.

Definition 2.2. A symmetric on a set X is a mapping d : X × X −→ [0, ∞) such that

1. d(x, y) = 0 if and only if x = y, and

2. d(x, y) = d(y, x).

A set X , together with a symmetric d is called a symmetric space.

3. P D -operator pair

Definition 3.1. Let f , g : X −→ X be mappings. The pair (f , g ) is said to be P D -operator pair if there is a point u in X such
that u ∈ C (f , g ) and d(fgu, gfu) ≤ diam(PC (f , g )).

Remark 3.1. (1) In the following, we can see that every pair of commuting, nontrivial weakly compatible and nontrivial
occasionally weakly compatible self mappings are P D -operator pair. In addition, P D -operator pairs are weaker form of
weakly commuting and compatible maps, but the reverse implications need not be true.

Commuting maps −−−−→ Weakly commuting maps −−−−→ Compatible maps

 
 
 
P D -operator pair ←−−−− Weakly compatible maps

↖ ↙
Occasionally weakly compatible maps

Example 3.1. Let X = [0, 1] and defined d : X × X −→ [0, ∞) by d(x, y) = (x − y)2 .

Define f , g : X −→ X by f (x) = x2 and g (x) = x/2, for all x ̸= 0 and f (0) = g (0) = 1. Since, C (f , g ) = {0, 1/2} and
PC (f , g ) = {1, 1/4}, diam(C (f , g )) = 1/4 and diam(PC (f , g )) = 9/16.
Clearly, (f , g ) is P D -operator pair, but not commuting, not weakly compatible and not occasionally weakly compatible.
 H.K. Pathak, Deepmala / Journal of Computational and Applied Mathematics 239 (2013) 103–113 105

Example 3.2. Let X = [0, 1] and defined d : X × X −→ [0, ∞) by

ex−y − 1,

if x ≥ y
d(x, y) =
ey−x , if x < y.
Define f , g : X −→ X by

1, if x = 0,

1, if x = 0,

f ( x) = x2 , if 0 < x ≤ 1/2, g (x) =
1, x/2, if 0 < x ≤ 1.
if 1/2 < x ≤ 1.

Here C (f , g ) = {0, 1/2} and PC (f , g ) = {1, 1/4}.

Clearly, (f , g ) is P D -operator pair, but not commuting, not weakly compatible and not occasionally weakly compatible.
(2) P D -operator is different form P -operator pair and JH -operator pair. For this, consider the following examples.

Example 3.3. Let X = [0, ∞) be endowed with the Euclidean metric, f , g : X −→ X be two mappings defined by, f (x) = 2x
and g (x) = 2x2 , for all x ̸= 0 and f (0) = g (0) = 3.
Since, C (f , g ) = {0, 1}, PC (f , g ) = {2, 3}, diam(C (f , g )) = 1 and diam(PC (f , g )) = 1.
Clearly, (f , g ) is P -operator and JH -operator pair, but not P D -operator pair.
Hence, every P -operator and JH -operator pair need not be a P D -operator pair.

Example 3.4. Let X = [0, ∞) and defined d : X × X −→ [0, ∞) by d(x, y) = e|y−x| . Define f , g : X −→ X by
2, if x = 0,

x + 2, if 0 ≤ x < 2,

f (x) = 3x, if 0 < x ≤ 1, g (x) =
x, if 2 ≤ x < ∞.
x/2, if 1 < x < ∞.

Clearly, f (0) = g (0) = 2. and f (1) = g (1) = 3. Here C (f , g ) = {0, 1} and PC (f , g ) = {2, 3}. diam(C (f , g )) = e1 and
diam(PC (f , g )) = e1 .
Clearly, (f , g ) is P D -operator pair, but neither P -operator and nor JH -operator pair.
Hence, every P D -operator pair need not be P -operator and JH -operator pair.

4. Fixed point theorems for P D -operator pairs under relaxed conditions

In this section, we prove some fixed point theorems for P D -operator pairs on the space (X , d), without imposing the
restriction of the triangle inequality or symmetry on d. Let φ : R+ −→ R+ be nondecreasing function satisfying the condition
φ(t ) < t, for each t > 0. We now prove the following theorem.
Theorem 4.1. Let X be a nonempty set and d : X × X −→ [0, ∞) be a function satisfying the condition (1). Suppose (f , g ) is
P D -operator pair and satisfy the condition:
d(fx, fy) ≤ φ(max{d(gx, gy), d(gx, fy), d(fx, gy), d(gy, fy)}), (2)
for each x, y ∈ X . Then f and g have a unique common fixed point.
Proof. Since, (f , g ) is P D -operator pair, there is a point u in X such that fu = gu and
d(fgu, gfu) ≤ diam(PC (f , g )). (3)
First, we prove that PC (f , g ) is singleton. Suppose w and z be two distinct points in X such that w = fu = gu and
z = f v = g v , for some u, v ∈ C (f , g ). Then from (2), we obtain
d(w, z ) = d(fu, f v) ≤ φ(max{d(gu, g v), d(gu, f v), d(fu, f v), d(g v, f v)})
≤ φ(max{d(fu, f v), d(fu, f v), d(fu, f v), d(f v, f v)}),
≤ φ(d(fu, f v)) < d(w, z ),
a contradiction. So, w = z i.e. w = fu = gu = f v = g v = z.
Thus, PC (f , g ) is singleton i.e., w = fu = gu is the unique point of coincidence.
diam(PC (f , g )) = 0. Form (3), fgu = gfu, for some points u ∈ C (f , g ). Now, by (2), we have
d(ffu, f v) = d(ffu, fu) ≤ φ(max{d(gfu, g v), d(gfu, f v), d(ffu, g v), d(g v, f v)})
≤ φ(max{d(ffu, fu), d(ffu, fu), d(ffu, fu), 0})
≤ φ(d(ffu, fu)),
a contradiction. Hence, ffu = gfu = fu, f and g have a common fixed point. Uniqueness, is obvious. So, f and g have a unique
common fixed point. This completes the proof of the theorem. 
106 H.K. Pathak, Deepmala / Journal of Computational and Applied Mathematics 239 (2013) 103–113

Corollary 4.2. Let X be a nonempty set and d : X × X −→ [0, ∞) be a function satisfying the condition (1). Suppose (f , g ) is a
P D -operator pair and satisfy the condition:
d(fx, fy) ≤ φ(d(gx, gy)), (4)

for each x, y ∈ X . Then f and g have a unique common fixed point.

Remark 4.3. Corollary 4.2 is a generalization of Corollary 2.1 of Bhatt et al. [7] for P D -operator pair.

The proof of the following theorem can be easily obtained by replacing condition (2) by condition (5) in the proof of
Theorem 4.1.

Theorem 4.4. Let X be a nonempty set and d : X × X −→ [0, ∞) be a function satisfying the condition (1). Suppose (f , g ) is a
P D -operator pair satisfying the condition
d(fx, fy) ≤ φ(max{d(gx, gy), d(gx, fy), d(fy, gy)}), (5)

for each x, y ∈ X . Then f and g have a unique common fixed point.

Remark 4.5. It is clear to note that in the proof of above result we need not to assume symmetry and triangle inequality of
d. Hence the above result is a proper generalization of the results due to Aamri and El Moutawakil [12] for P D -operator
pair.

Theorem 4.6. Let X be a nonempty set and d : X × X −→ [0, ∞) be a function satisfying the condition (1). Suppose that f , g,
S and T are self mappings of X and that (f , S ) and (g , T ) are each P D -operators and satisfy the condition:

d(fx, gy) ≤ φ(max{d(Sx, Ty), d(Sx, gy), d(Ty, gy)}), (6)

for each x, y ∈ X . Then f , g, S and T have a unique common fixed point.

Proof. Since the pair (f , S ) and (g , T ) are each P D -operators, there exists points x, y ∈ X such that fx = Sx and gy = Ty.
Also, d(fSx, Sfx) ≤ diam(PC (f , S )) and d(gTy, Tgy) ≤ diam(PC (g , T )). We claim that fx = gy, otherwise, by (6), we have

d(fx, gy) ≤ φ(max{d(Sx, Ty), d(Sx, gy), d(Ty, gy)})

d(fx, gy) ≤ φ(max{d(fx, gy), d(fx, gy), d(gy, gy)})
< d(fx, gy),
a contradiction. Thus fx = Sx = gy = Ty. Moreover, if there is another point u ∈ X such that fu = Su, then using (6), it
follows that fu = Su = gy = Ty, or fx = fu. Hence x is the unique point of coincidence of f and S. Similarly, y is the unique
point of coincidence of g and T . Thus, diam(PC (f , S )) = 0 and diam(PC (g , T )) = 0. Hence fSx = Sfx, gTy = Tgy. By condition
(6), we get

d(ffx, gy) = d(ffx, fx) ≤ φ(max{d(Sfx, Ty), d(Sfx, gy), d(Ty, gy)})
≤ φ(max{d(ffx, fx), d(ffx, fx), 0})
≤ φ(d(ffx, fx)) < d(ffx, fx),
a contradiction. Thus, ffx = Sfx = fx, i.e. f and S have a common fixed point, which is obviously unique. By similar argument,
we can show that g and T have a unique common fixed point. For uniqueness, suppose there are points v, w ∈ X and v ̸= w
such that v = f v = S v and w = g w = T w . By (6), we get

d(v, w) = d(f v, g w) < d(v, w),

a contradiction. Hence, f , g, S and T have a unique common fixed point. This completes the proof of the theorem. 

Remark 4.7. The above result is a proper extension and generalization of the results due to [11,12] respectively, under
relaxed conditions without assuming completeness or weak compatibility on X or symmetry or triangle inequality on d for
P D -operator pairs.
The proof of the following theorem can be easily obtained by replacing condition (2) by condition (7) in the proof of
Theorem 4.1.

Theorem 4.8. Let X be a nonempty set and d : X × X −→ [0, ∞) be a function satisfying the condition (1). Suppose (f , g ) is
P D -operator pair and satisfy the condition:
d(fx, fy) < max{d(gx, gy), d(gx, fy), d(gy, fx), d(gy, fy)}, (7)

for each x, y ∈ X . Then f and g have a unique common fixed point.

 H.K. Pathak, Deepmala / Journal of Computational and Applied Mathematics 239 (2013) 103–113 107

Remark 4.9. Theorem 4.8 is a proper generalization of result due to Bhatt et al. [7].

Example 4.1. Let X = [0, 1] and defined d : X × X −→ [0, ∞) by

ex−y − 1,

if x ≥ y
d(x, y) =
ey−x , if x < y.

Define f , g : X −→ X by f (x) = x/3 and g (x) = x/2 for all x, y ∈ X . Here C (f , g ) = {0}.
By a routine calculation one can show that the contractive condition (7) holds for every x ̸= y ∈ X . Also, notice that (f , g ) is
a P D -operator pair. Therefore, all the the conditions of Theorem 4.8 are satisfied and 0 is the common unique fixed point
of f and g.

Theorem 4.10. Let X be a nonempty set and d : X × X −→ [0, ∞) be a function satisfying the condition (1). Suppose that f
and g be a selfmaps on X . Suppose (f , g ) is a P D -operator pair and satisfy the condition:

d(fx, fy) ≤ a d(gx, gy) + b max{d(fx, gx), d(fy, gy)} + c max{d(gx, gy), d(gx, fx), d(gy, fy)}, (8)
√
for each x, y ∈ X where a, b, c > 0, a + b + c = 1 and a + c < a. Then f and g have a unique common fixed point.

Proof. By hypothesis, there is a point u ∈ X such that fu = gu. Suppose that there exists another point v ∈ X for which
f v = g v . Then from (8),

d(fu, f v) ≤ a d(fu, f v) + b max{0, 0} + c max{d(fu, f v), 0, 0}

= (a + c ) d(fu, f v).
Since, a + c < 1, the above inequality implies that d(fu, f v) = 0, which in turn implies that fu = f v . Therefore fu is unique,
diam(PC (f , g )) = 0 and fgu = gfu, for some points u ∈ C (f , g ). Now, by (8), we get d(ffu, f v) = d(ffu, fu) < (d(ffu, fu)), a
contradiction.
Hence, ffu = gfu = fu, f and g have a common fixed point. Uniqueness, is obvious. So, f and g have a unique common fixed
point. This completes the proof of the theorem. 

Remark 4.11. The above result is a proper generalization of Theorem 2 of Jungck and Rhoades [8] and Theorem 2.3 of Bhatt
et al. [7] for P D -operator pair.

Let the control function Φ : R+ −→ R+ be a continuous nondecreasing function such that Φ (2t ) ≤ 2Φ (t ) and, Φ (0) = 0
iff t = 0.
Let ψ : R+ −→ R+ be another function such that ψ(t ) < t, for each t > 0.

Theorem 4.12. Let X be a nonempty set and d : X × X −→ [0, ∞) be a function satisfying the condition (1). Suppose (f , S )
and (g , T ) are P D -operator pairs and satisfy the condition:

Φ (d(fx, gy)) ≤ ψ(MΦ (x, y)), where (9)

MΦ (x, y) = max {Φ (d(Sx, Ty)), Φ (d(Sx, fx)), Φ (d(gy, Ty)), 1/2[Φ (d(fx, Ty)) + Φ (d(Sx, gy))]}, (10)

for each x, y ∈ X . Then f , g, S and T have a unique common fixed point.

Proof. Since the pair (f , S ) and (g , T ) are P D -operator pairs, there exists points x, y ∈ X such that fx = Sx and gy = Ty.
Also, d(fSx, Sfx) ≤ diam(PC (f , S )) and d(gTy, Tgy) ≤ diam (PC (g , T )). We claim that fx = gy, otherwise, by (9) and (10), we
have

0 < Φ (d(fx, gy)) ≤ ψ(MΦ (d(fx, gy)))

= ψ(Φ (d(fx, gy)))
< Φ (d(fx, gy)),
a contradiction. Thus fx = Sx = gy = Ty. Moreover, if there is another point u ∈ X such that fu = Su, then using (9) and
(10), it follows that fu = Su = gy = Ty, or fx = fu. Hence x is the unique point of coincidence of f and S. Similarly, y is the
unique point of coincidence of g and T . Thus, diam(PC (f , S )) = 0 and diam(PC (g , T )) = 0. Hence fSx = Sfx, gTy = Tgy.
Hence from the repeated use of condition (9) and (10), we can show that f , g, S and T have a unique common fixed point.
This completes the proof of the theorem. 

Remark 4.13. The above Theorem 4.12 is an extension of result of Hussain et al. [6] from JH -operator to P D -operator
pairs and a proper generalization of results due to Bhatt et al. [7] and Jungck and Rhoades [8] for the pair of P D -operators
without assuming symmetry or triangle inequality.
108 H.K. Pathak, Deepmala / Journal of Computational and Applied Mathematics 239 (2013) 103–113

5. Fixed point theorems for P D -operator pairs in symmetric spaces

In this section, we prove some fixed point theorems for P D -operator pairs on the space (X , d), without imposing the
restriction of the triangle inequality and assuming symmetry on d.

Theorem 5.1. Let f and g be selfmaps of symmetric space X . Suppose (f , g ) is a P D -operator pair and satisfy the condition:
d(fx, fy) ≤ h max{d(gx, gy), d(fx, gx), d(fy, gy), d(fx, gy), d(fy, gx)}, (11)
for each x, y ∈ X , 0 ≤ h < 1. Then f and g have a unique common fixed point.
Proof. Since, (f , g ) is P D -operator pair, there exists some points u ∈ C (f , g ) such that
d(fgu, gfu) ≤ diam(PC (f , g )). (12)
First, we prove that PC (f , g ) is singleton. Suppose not, let w, z ∈ PC (f , g ) such that w = fu = gu and z = f v = g v ,
∀u, v, w, z ∈ X . Let w ̸= z, then from (11), we have
d(w, z ) = d(fu, f v) ≤ h(max{d(gu, g v), d(fu, gu), d(f v, g v), d(fu, g v), d(f v, gu)})
≤ h (d(fu, f v))
< d(w, z ),
a contradiction. So, w = z i.e. w = fu = gu = f v = g v = z.
Thus, PC (f , g ) is singleton, i.e. w = fu = gu is the unique point of coincidence.
diam(PC (f , g )) = 0. Form (12), fgu = gfu, for some points u ∈ C (f , g ). Now, by (11), we have
d(ffu, f v) = d(ffu, fu) ≤ h(max{d(gfu, g v), d(ffu, gfu), d(f v, g v), d(ffu, g v), d(f v, gfu)})
≤ h (max{d(ffu, fu), d(fu, ffu)})
< d(ffu, fu),
a contradiction. Hence, ffu = gfu = fu, f and g have a common fixed point. Uniqueness, is obvious. So, f and g have a unique
common fixed point. This completes the proof of the theorem. 

Remark 5.2. Theorem 5.1 contains Theorem 2.1 of Al-Thagafi and Shahzad [4], Pathak and Hussain [5] and Jungck and
Hussain [9] as a special case.

Corollary 5.3. Let f and g be selfmaps of symmetric space X . Suppose (f , g ) is a P D -operator pair and satisfy the condition:
d(fx, fy) ≤ h max{d(gx, gy), d(gx, fx), d(gy, fy)}, (13)
for each x, y ∈ X , 0 ≤ h < 1. Then f and g have a unique common fixed point.
The proof of the following theorem can be easily obtained by replacing condition (11) by condition (14) in the proof of
Theorem 5.1.

Theorem 5.4. Let f and g be a selfmaps of symmetric space X . Suppose (f , g ) is a P D -operator pair and satisfy the condition:
d(fx, fy) ≤ φ(max{d(gx, gy), d(gx, fy), d(gy, fx), d(gy, fy)}), (14)
for each x, y ∈ X , where φ is defined as in Section 4. Then f and g have a unique common fixed point.
The proof of the following theorem can be easily obtained by replacing condition (11) by condition (7) in the proof of
Theorem 5.1.

Theorem 5.5. Let f and g be a selfmaps of symmetric space X . Suppose (f , g ) is a P D -operator pair and satisfy the condition (7).
Then f and g have a unique common fixed point.

Remark 5.6. Theorems 5.4 and 5.5 are proper generalization of the results due to Bhatt et al. [7] for P D -operator pair and
an extension of results due to Hussain et al. [6] from JH -operator to P D -operator pair.
The proof of the following theorem can be easily obtained by replacing condition (11) by condition (15) in the proof of
Theorem 5.1.

Theorem 5.7. Let f and g be a selfmaps of symmetric space X . Suppose (f , g ) is a P D -operator pair and satisfy the condition:
d(fx, fy) ≤ h max{d(gx, gy), (d(fx, gx) + d(fy, gy))/2, (d(fx, gy) + d(fy, gx))/2}, (15)
for each x, y ∈ X , 0 ≤ h < 1. Then f and g have a unique common fixed point.
 H.K. Pathak, Deepmala / Journal of Computational and Applied Mathematics 239 (2013) 103–113 109

Remark 5.8. The above result is a proper extension of the results due to Aamri and El Moutawakil [11] for P D -operator
pair and a special case of results of Imdad et al. [10] for P D -operator pair under relaxed conditions with out assuming
completeness or weak compatibility or E.A. property or symmetry or triangle inequality on d.

Theorem 5.9. Suppose that f , g, S and T are selfmaps of symmetric space X and that (f , S ) and (g , T ) are P D -operator pairs
and satisfy the condition:
d(fx, gy) ≤ φ(max{d(Sx, Ty), d(fx, Sx), d(gy, Ty), d(fx, Ty), d(gy, Sx)}), (16)
for each x, y ∈ X . Then f , g, S and T have a unique common fixed point.
Proof. Since the pair (f , S ) and (g , T ) are P D -operator pairs, there exists points x, y ∈ X such that fx = Sx and gy = Ty.
Also, d(fSx, Sfx) ≤ diam(PC (f , S )) and d(gTy, Tgy) ≤ diam(PC (g , T )). We claim that fx = gy, otherwise, by (16), we have
d(fx, gy) ≤ φ(max{d(fx, gy), d(fx, fx), d(gy, gy), d(fx, gy), d(gy, fx)})
d(fx, gy) ≤ φ(max{d(fx, gy), d(gy, fx)})
< d(fx, gy),
a contradiction. Thus fx = Sx = gy = Ty. Moreover, if there is another point u ∈ X such that fu = Su, then using (16), it
follows that fu = Su = gy = Ty, or fx = fu. Hence x is the unique point of coincidence of f and S. Similarly, y is the unique
point of coincidence of g and T . Thus, diam(PC (f , S )) = 0 and diam (PC (g , T )) = 0. Hence fSx = Sfx, gTy = Tgy.
By condition (16), we have
d(ffx, gy) = d(ffx, fx) < d(ffx, fx),
a contradiction. Thus, ffx = Sfx = fx, i.e. f and S have a common fixed point, which is obviously unique. By similar argument,
we can show that g and T have a unique common fixed point. For uniqueness, suppose there are points v, w ∈ X and v ̸= w
such that v = f v = S v and w = g w = T w . By (16), we have d(v, w) = d(f v, g w) < d(v, w), a contradiction. Hence, f , g,
S and T have a unique common fixed point. This completes the proof of the theorem. 

Remark 5.10. A special case of Theorem 5.9 is a proper generalization of Theorems 4 and 5 of Jungck and Rhoades [8] for
P D -operator pair.
The proof of the following theorem can be easily obtained by replacing condition (16) by condition (17) in the proof of
Theorem 5.9.

Theorem 5.11. Suppose that f , g, S and T are selfmaps of symmetric space X and that (f , S ) and (g , T ) are P D -operator pairs
and satisfy the condition:
d(fx, gy) ≤ φ(max{d(Sx, Ty), d(Sx, fx), d(Ty, gy), d(Sx, gy), d(Ty, fx)}), (17)
for each x, y ∈ X . Then f , g, S and T have a unique common fixed point.

Remark 5.12. The above result is a proper generalization of results due to Jungck and Rhoades [8] and Bhatt et al. [7].

Example 5.1. Let X = [0, 1] and defined d : X × X −→ [0, ∞) by d(x, y) = (x − y)2 . Define f , g , S , T : X −→ X by
f (x) = g (x) = x/2 and S (x) = T (x) = x, for all x, y ∈ X .
Let φ : R+ −→ R+ such that φ(t ) = t /2, for t > 0. Clearly, φ(t ) < t is nondecreasing function. In order, to verify contractive
condition (17), we see that
If x = 0, 0 < y ≤ 1, then
d(fx, gy) = d(0, y/2) = (y/2)2 < y2 /2 = φ(y2 ) ≤ φ(d(Sx, Ty)).
If 0 < x < y ≤ 1, then
d(fx, gy) = d(x/2, y/2) = ((x − y)/2)2 < (x − y)2 /2 = φ((x − y)2 ) ≤ φ(d(Sx, Ty)).
By symmetry of d, other cases follows. Thus, all the conditions of Theorem 5.11 are satisfied and 0 is the unique common
fixed point of f , g, S and T .
Now, we prove the following theorem-

Theorem 5.13. Suppose that f , g, S and T are selfmaps of symmetric space X and that (f , S ) and (g , T ) are P D -operator pairs
and satisfy the condition:
(d(fx, gy))p < ad(fx, Ty)p + (1 − a) max {(d(fx, Sx))p , (d(gy, Ty))p , (d(fx, Sx))p/2 (d(fx, Ty))p/2 ,
(d(Ty, fx))p/2 (d(Sx, gy))p/2 }, (18)
for each x, y ∈ X , where 0 < a, and p ≥ 1. Then f , g, S and T have a unique common fixed point.
110 H.K. Pathak, Deepmala / Journal of Computational and Applied Mathematics 239 (2013) 103–113

Proof. Since the pair (f , S ) and (g , T ) are P D -operator pairs, there exists points x, y ∈ X such that fx = Sx and gy = Ty.
Also, d(fSx, Sfx) ≤ diam(PC (f , S )) and d(gTy, Tgy) ≤ diam(PC (g , T )). We claim that fx = gy, otherwise, by (18), we have

(d(fx, gy))p < ad(fx, gy)p + (1 − a) max {0, 0, 0, (d(gy, fx))p/2 (d(fx, gy))p/2 }
(d(fx, gy))p < d(fx, gy)p ,
a contradiction. Thus fx = Sx = gy = Ty. Moreover, if there is another point u ∈ X such that fu = Su, then using (18), it
follows that fu = Su = gy = Ty, or fx = fu. Hence x is the unique point of coincidence of f and S. Similarly, y is the unique
point of coincidence of g and T . Thus, diam(PC (f , S )) = 0 and diam(PC (g , T )) = 0. Hence fSx = Sfx, gTy = Tgy.
Now from the repeated use of (18), we can show that f , g, S and T have a unique common fixed point. This completes the
proof of the theorem. 

Remark 5.14. The above Theorem 5.13 is an extension of result of Hussain et al. [6] and a proper generalization of results
due to Bhatt et al. [7] and Jungck and Rhoades [8] for P D -operator pair.

Remark 5.15. Since, as we see that every commuting, nontrivial weakly compatible and occasionally weakly compatible
mappings are P D -operator pairs, but converse need not be true. Hence, one can generalize many results of commuting,
weakly compatible and owc maps for different contractive conditions by using the concept of P D -operator pair.

6. Application in dynamic programming

Throughout this section, we assume that U and V are Banach spaces, W ⊆ U and D ⊆ V . Let R denote the fields of reals,
τ : W × D −→ W , l, m : W × D −→ R and G, F : W × D × R −→ R. Consider W and D as the state and decision spaces
respectively, the problem of dynamic programming reduces to the problem of solving the functional equations:

p(x) = max{l(x, y) + G(x, y, p(τ (x, y)))}, x ∈ W. (19)

y∈D

q(x) = max{m(x, y) + F (x, y, q(τ (x, y)))}, x ∈ W. (20)

y∈D

In the multistage process, some functional equations arise in a natural way [14]. In this section, we study the existence of
common solutions of the functional Eqs. (19) and (20) arising in dynamic programming.
Let B(W ) denotes the set of all bounded real-valued functions on W . For an arbitrary h ∈ B(W ), define ∥h∥ = supx∈W |h(x)|.
Then (B(W ), ∥ · ∥) is a Banach space.

Theorem 6.1. Suppose that the following conditions (i)–(iii) are satisfied:
(i) G, F , l and m are bounded.
(ii) Assume that for every (x, y) ∈ W × D and z , z0 ∈ R
|G(x, y, z ) − G(x, y, z0 )| ≤ φ(|g (z ) − g (z0 )|), (21)
where φ : R +
−→ R is a nondecreasing function satisfying the condition φ(t ) < t for each t > 0, f and g are defined as
+

follows:
fh(x) = max{l(x, y) + G(x, y, h(τ (x, y)))}, x ∈ W , h ∈ B(W ). (22)
y∈D

gk(x) = max{m(x, y) + F (x, y, k(τ (x, y)))}, x ∈ W , k ∈ B(W ). (23)

y∈D

(iii) d(fgu(x), gfu(x)) ≤ diam(PC (f , g )), for some u(x) ∈ C (f , g ), where u(x) ∈ B(W ).
Then the functional Eqs. (19) and (20) has a unique bounded solution.

Proof. For any h, k ∈ B(W ), let

d(h, k) = sup{|h(x) − k(x)| : x ∈ W }.

From conditions (i) to (iii), it follows that f and g are self mappings of B(W ).
Let h1 , h2 ∈ B(W ), x ∈ W and η > 0, then there exists y1 , y2 ∈ D such that

fh1 (x) < l(x, y1 ) + G(x, y1 , h1 (x1 )) + η, (24)

fh2 (x) < l(x, y2 ) + G(x, y2 , h2 (x2 )) + η, (25)
fh1 (x) ≥ l(x, y2 ) + G(x, y2 , h1 (x2 )), (26)
fh2 (x) ≥ l(x, y1 ) + G(x, y1 , h2 (x1 )), (27)
 H.K. Pathak, Deepmala / Journal of Computational and Applied Mathematics 239 (2013) 103–113 111

where xi ∈ τ (x, yi ), i = (1, 2). Subtracting (27) from (24) and using (ii), we have

fh1 (x) − fh2 (x) < G(x, y1 , h1 (x1 )) − G(x, y1 , h2 (x1 )) + η

≤ |G(x, y1 , h1 (x1 )) − G(x, y1 , h2 (x1 ))| + η
≤ φ(|gh1 (x1 ) − gh2 (x1 )|) + η
≤ φ(d(gh1 , gh2 )) + η.
Similarly, from (25) and (26), we get

fh1 (x) − fh2 (x) > φ(d(gh1 , gh2 )) − η.

Hence,

|fh1 (x) − fh2 (x)| < φ(d(gh1 , gh2 )) + η. (28)

Since (28) is true for any x ∈ W and η is any positive number,

|fh1 (x) − fh2 (x)| < φ(d(gh1 , gh2 )). (29)

Therefore, from Corollary 4.2, there is the unique common fixed point of f and g, i.e., there is a unique bounded solution of
functional Eqs. (19) and (20). This completes the proof of the theorem. 

Example 6.1. Let U = V = R = (−∞, +∞), W = [1, 2]∪{0} and D = [1, 10]. Then Theorem 6.1 ensure that the following
functional equations:
 
x
p(x) = max 2xy + sin (p(2x + 3y)) , x∈W (30)
y∈D 2 + 2xy
xy5
 
q(x) = max 3xy2 q(sin (x + y)) + q(sin (x + y)) , x∈W (31)
y∈D 1 + xy

possesses a unique bounded solution in B(W ).

To see this, let φ : R+ −→ R+ be defined by φ(t ) = t /2, for each t > 0 and for each h, k ∈ B(W ), f and g are defined as
follows,
 
x
fh(x) = max 2xy + sin (h(2x + 3y)) , x ∈ W.
y∈D 2 + 2xy
xy5
  
gk(x) = max 3xy2 + k(sin (x + y)) , x ∈ W.
y∈D 1 + xy
In fact, taking

l(x, y) = 2xy, m(x, y) = 3xy2 k(sin (x + y)),

x xy5
G(x, y, z ) = sin z , and F (x, y, z0 ) = z0 .
2 + 2xy 1 + xy
Clearly, G, F , l and m are bounded and

|G(x, y, z ) − G(x, y, z0 )| ≤ φ(|g (z ) − g (z0 )|), x ∈ W , y ∈ D, z , z0 ∈ R.

Also, (f , g ) is a P D -operator pair, since

d(fg (0), gf (0)) = 0 ≤ diam(PC (f , g )), for 0 ∈ C (f , g ).

Thus, all the conditions of Theorem 6.1 are satisfied, so it guarantees that the functional Eqs. (30) and (31) possesses a unique
bounded solution in B(W ).

7. Application in variational inequalities

In this section, we apply Corollary 5.3 to show the existence of solutions of variational inequalities as in [15]. Variational
inequalities arise in optimal stochastic control and other problems in mathematical physics, for examples, deformation of
elastic bodies stretched over solid obstacles, elasto-plastic torsion etc. [16]. The iterative method for solutions of discrete
variational inequalities are very suitable for implementation on parallel computers with single instruction, multiple-data
architecture, particularly on massively parallel processors.
112 H.K. Pathak, Deepmala / Journal of Computational and Applied Mathematics 239 (2013) 103–113

The variational inequality problem is to find a function u such that

max {Lu − f , u − φ} = 0 on Ω ,

(32)
u=0 on ∂ Ω ,

where Ω is a bounded open convex subset of RN with smooth boundary, L is an elliptic operator defined on Ω by
L = −aij (x)∂ 2 /∂ xi ∂ xj + bi (x)∂/∂ xi + c (x)IN ,
where summation with respect to repeated indices is implied, c (x) ≥ 0, [aij (x)] is a strictly positive definite matrix, uniformly
in x, for x ∈ Ω , f and φ are smooth functions defined in Ω and f satisfies the condition: φ(x) ≥ 0 for x ∈ Ω .
A problem related to (32) is the two-obstacle variational inequality. Given two functions φ and u defined on Ω such that
φ ≤ µ in Ω , φ ≤ 0 ≤ µ on ∂ Ω , the corresponding variational inequality is as follows:
max{min [(Lu − f , u − φ), u − µ]} = 0 on Ω ,

(33)
u=0 on ∂ Ω .
The problem (33) arises in stochastic game theory.
Let A be an N × N matrix corresponding to the finite difference discretizations of the operator L. We shall make the following
assumptions about the matrix A:

Aii = 1, Aij > −1, Aij < 0 for i ̸= j. (34)
j;j̸=i

These assumptions are related to the definition of M-matrices; matrices arising from the finite difference discretization of
continuous elliptic operators under the appropriate conditions and Q denotes the set of all discretized vectors.
Let B = IN − A. Then the corresponding property for the matrix B will be

Bii = 0, Bij < 1, Bij > 0 for i ̸= j. (35)
j;j̸=i

Bij and A∗ be an N × N matrix such that A∗ii = 1 − q and A∗ij = −q for i ̸= j. Then we have B∗ = IN − A∗ .

Let q = maxi j
Now, we show the existence of iterative solutions of variational inequalities: Consider the following simultaneous discrete
variational inequality mentioned above:

max [min {A(x − A∗ d(Ix, Tx)) − f , x − A∗ d(Ix, Tx) − φ}, x − A∗ d(Ix, Tx) − µ] = 0 (36)
where I is one–one, (I , T ) is a P D -operator pair from R into itself implicitly defined by:
N

Tx = min k[max {BIx + A(1 − B∗ ) d(Ix, Tx) + f , (1 − B∗ ) d(Ix, Tx) + φ}, (1 − B∗ ) d(Ix, Tx) + µ], (37)
for all x ∈ Q , 0 < k < 1. Then (36) is equivalent to the common fixed point problem:
x = Ix = Tx. (38)
In a two-person game we determine the best strategies for each player on the basis of maximin and minimax criterion
of optimality. This criterion will be well stated as follows: A player lists his/her worst possible outcomes and then he/she
chooses that strategy which corresponds to the best of these worst outcomes. Here, the problem (36) exhibits the situation
in which two players are trying to control a diffusion process; the first player is trying to maximize a cost functional, and
the second player is trying to minimize a similar functional. The first player is called the maximizing player and the second
one the minimizing player. Here, f represents the continuous rate of cost for both players, φ is the stopping cost for the
maximizing player, and µ is the stopping cost for the minimizing player. This problem is fixed by inducting a P D -operator
pair (T , I ) under conditions as stated above.

Theorem 7.1. Under the assumption (34) and (35) a solution of (38) exists.
Proof. Let (Ty)i = k[(1 − B∗ij ) d(Iyj , Tyj ) + µi ], for any y ∈ Q and any i, j = 1, 2, . . . , N.
Now for any x ∈ Q , since (Tx)i ≤ k[(1 − B∗ij ) d(Ixj , Txj ) + µi ], we have

(Tx)i − (Ty)i ≤ k (1 − B∗ij ) max {d(Ixj , Txj ), d(Iyj , Tyj )}. (39)
If (Ty)i = max k {Bij Iyj + (1 − Bij )d(Iyj , Tyj ) + fi , (1 − Bij )d(Iyj , Tyj ) + φi }, i.e. if the maximizing player succeeds to maximize
∗ ∗

a cost functional in his/her strategy which corresponds to the best of N worst outcomes from his/her list, then the game
would be one sided. In this situation, we introduce the one sided operator:
T + x = max k{BIx + A(1 − B∗ ) d(Ix, Tx) + f , (1 − B∗ ) d(Ix, Tx) + φ}.
Therefore, we have (Ty)i = (T + y)i .
 H.K. Pathak, Deepmala / Journal of Computational and Applied Mathematics 239 (2013) 103–113 113

Now, since (Tx)i ≤ (T + x)i , we have

(Tx)i − (Ty)i ≤ (T + x)i − (T + y)i . (40)

Now, if (Tx)i = k [Bij Ixj + Aij (1 − Bij )d(Ixj , Txj ) + fi ], then since (Ty)i ≥ k [Bij Iyj + Aij (1 − Bij )d(Iyj , Tyj ) + fi ], by using (34), we
∗ ∗

have
(T + x)i − (T + y)i ≤ k [Bij d(Ixi , Tyi ) + (1 − B∗ij ) max{d(Ixj , Txj ), d(Iyj , Tyj )}]. (41)
If (Tx)i = k[(1 − B∗ij )d(Ixj , Txj ) + φi ], then since (Ty)i ≥ k[(1 − B∗ij )d(Iyj , Tyj ) + φi ], we have

(Tx)i − (Ty)i ≤ k (1 − B∗ij ) max {d(Ixj , Txj ), d(Iyj , Tyj )}.

Hence, from (34) to (36), we have
(Tx)i − (Ty)i ≤ k[q d(Ix, Iy) + (1 − q) max {d(Ix, Tx), d(Iy, Ty)}]. (42)
Since x and y are arbitrarily chosen, we have
(Ty)i − (Tx)i ≤ k[q d(Ix, Iy) + (1 − q) max {d(Ix, Tx), d(Iy, Ty)}]. (43)
Therefore, from (42) and (43), it follows that
d(Tx, Ty) ≤ k max{d(Ix, Iy), d(Ix, Tx), d(Ty, Ty)}. (44)
Hence, we see that all the conditions of Corollary 5.3 are satisfied. Therefore, Corollary 5.3 ensure the existence of a solution
of (38) exists. 

Acknowledgments

The authors are grateful to the referee for his/her valuable comments and suggestions for the improvement of the paper.
The research of the second author is supported by the Department of Science and Technology, Government of India under
INSPIRE Fellowship Program No. DST/INSPIRE Fellowship/2010/[182].

References

[1] G. Jungck, Commuting mappings and fixed points, Amer. Math. Monthly 83 (1976) 261–263.
[2] S. Sessa, On a weak commutativity condition of mappings in fixed point consideration, Publ. Inst. Math. 32 (1982) 149–153.
[3] G. Jungck, Compatible mappings and common fixed points, Int. J. Math. Math. Sci. 9 (1986) 771–779.
[4] M.A. Al-Thagafi, N. Shahzad, Generalized I-nonexpansive selfmaps and invariant approximations, Acta Math. Sin. (Engl. Ser.) 24 (5) (2008) 867–876.
[5] H.K. Pathak, N. Hussain, Common fixed points for P -operator pair with applications, Appl. Math. Comput. 217 (2010) 3137–3143.
[6] N. Hussain, M.A. Khamsi, A. Latif, Common fixed points for JH -operators and occasionally weakly biased pairs under relaxed conditions, Nonlinear
Anal. 74 (2011) 2133–2140.
[7] A. Bhatt, H. Chandra, D.R. Sahu, Common fixed point theorems for occasionally weakly compatible mappings under relaxed conditions, Nonlinear
Anal. 73 (2010) 176–182.
[8] G. Jungck, B.E. Rhoades, Fixed point theorems for occasionally weakly compatible mappings, Fixed Point Theory 7 (2) (2006) 287–296.
[9] G. Jungck, N. Hussain, Compatible maps and invariant approximations, J. Math. Anal. Appl. 325 (2007) 1003–1012.
[10] M. Imdad, Javid Ali, L. Khan, Coincidence and fixed points in symmetric spaces under strict contractions, J. Math. Anal. Appl. 320 (2006) 352–360.
[11] M. Aamri, D. El Moutawakil, Some new common fixed point theorems under strict contractive conditions, J. Math. Anal. Appl. 270 (2002) 181–188.
[12] M. Aamri, D. El Moutawakil, Common fixed points under contractive conditions in symmetric spaces, Appl. Math. E-Notes 3 (2003) 156–162.
[13] G. Jungck, B.E. Rhoades, Fixed points for set valued functions with out continuity, Indian J. Pure Appl. Math. 29 (3) (1998) 227–238.
[14] R. Bellman, E.S. Lee, Functional equations arising in dynamic programming, Aequationes Math. 17 (1978) 1–18.
[15] S.A. Belbas, I.D. Mayergoyz, Application of fixed point methods, Numer. Math. 51 (1987) 631–654.
[16] G. Duvaut, J.L. Lions, Inequalities in Mechanics and Physics, Springer, Berlin, 1976.