Proyecciones Journal of Mathematics

Vol. 35, No 4, pp. 457-468, December 2016.

Universidad Católica del Norte
Antofagasta - Chile

Stability of generalized Jensen functional

equation on a set of measure zero

Hajira Dimou
University of Ibn Tofail, Morocco
Youssef Aribou
University of Ibn Tofail, Morocco
Abdellatif Chahbi
University of Ibn Tofail, Morocco
and
Samir Kabbaj
University of Ibn Tofail, Morocco
Received : June 2016. Accepted : August 2016

Abstract
Let E is a complex vector space and F is real (or complex ) Ba-
nach space. In this paper, we prove the Hyers-Ulam stability for the
generalized Jensen functional equation
m−1
X
f (x + bk y) = mf (x), x, y ∈ E,
k=0

where f : E → F and bk = exp( 2iπk

m ) for 0 ≤ k ≤ m − 1, on a set of
Lebesgue measure 0.

Subjclass [2010] : Primary 39B82; Secondary 39B52.

Keywords : K- Jensen functional equation, Hyers-Ulam stability.

458 H. Dimou, Y. Aribou, A. Chahbi and S. Kabbaj

1. Introduction

The question concerning the stability of functional equations has been posed
by S. M. Ulam in 1940 [29], and its solution given by H.D. Hyers [13] in 1941.
Hyers theorem was generalized by T. Aoki [1] for additive mappings and
by Th. M. Rassias [22] for linear mappings by considering an unbounded
Cauchy difference. The paper [22] of Th. M. Rassias has provided a lot
of influence in the development of what we now call Hyers-Ulam-Rassias
stability of functional equations. During the last decades, many stability
problems for various functional equations have been studied by numerous
mathematicians, we refer, for example, to [12],[14],[15],[17].

The stability problems of several functional equations on a restricted

domain have been extensively investigated by a number of authors, for
example [[4], [9], [8], [10] [11], [17], [20], [26], [23] ].
It is very natural to ask if the restricted domain D = {(x, y) ∈ E2 :
||x|| + ||y|| ≥ d} can be replaced by a much smaller subset Ω ⊂ D (e.g. a
subset of measure 0 in a measure space E).
In 2013, J. Chung in [8] answered to this question by considering the
stability of the Cauchy functional equation

(1.1) f (x + y) = f (x) + f (y)

in a set Ω ⊂ {(x, y) ∈ R2 : ||x||+||y|| ≥ d} where m(Ω) = 0 and f : R → R.

In 2014, J. Chung and J. M. Rassias [9] proved the stability of the
quadratic functional equation in a set of measure zero.
Throughout we assume that E is C vector space and F vector space over
K = R or C and 2 ≤ m ∈ N. Our aim is to prove the Hyers-Ulam stability
for the generalized Jensen functional equation

m−1
X
(1.2) f (x + bk y) = mf (x), x, y ∈ E,
k=0

where bk = exp( 2iπk

m ) for 0 ≤ k ≤ m−1, on a set Lebesgue measure 0. These
results are applied to study of an asymptotic behavior of this functional
equation. The stability and solution of this equation and its generalizations
were studied by numerous researchers, see for example [5, 2, 19, 28].
We also recall the following theorem which will be used in the sequel.
 Stability of generalized Jensen functional equation on a set of ... 459

Theorem 1.1 ([6]). Suppose that E is a complex vector space and F is

real (or complex) Banach space, m ≥ 2 ∈ N, bk = exp( 2ikπ
m ) for 0 ≤ k ≤
m − 1 and δ ≥ 0. If f : E → F satisfies
°m−1 °
°X °
° °
(1.3) ° f (x + bk y) − mf (x)° ≤ δ
° °
k=0

for all x, y ∈ E. Then there exists a unique generalized polynomial (GP)

q : E → F at most m − 1 such that q(0) = 0 and
kf (x) − f (0) − q(x)k ≤ 3m−1 δ f or all x ∈ E.
Moreover
m−1
X
q(x + bk y) − mq(x) = 0 for all x, y ∈ E.
k=0

2. Main result

For given x, y, t ∈ E
Px,y,t = {(x; t), (x + bk t; y), (x + bk y; t), k = 0, . . . , m − 1}.
Let Ω ⊂ E 2 . Throughout this section, we assume that Ω satisfies the
condition:
For given x, y ∈ E there exists t ∈ E such that

(C) {(x, t), (x + bk t, y), (x + bk y, t)} ⊂ Ω.

We prove the following stability theorem in Ω.
Theorem 2.1. Let δ ≥ 0. Suppose f : E → F satisfies the functional
inequality °m−1 °
°X °
° °
(2.1) ° f (x + bk y) − mf (x)° ≤ δ
° °
k=0

for all (x, y) ∈ Ω, where bk = exp( 2ikπ

m ), 0 ≤ k ≤ m − 1. Then there
exists a unique generalized polynomial (GP) q : E → F of degree at most
m − 1 such that q(0) = 0 and
(2.2) kf (x) − f (0) − q(x)k ≤ 3m δ f or all x ∈ E.
Moreover
m−1
X
q(x + bk y) − mq(x) = 0 for all x, y ∈ E.
k=0
460 H. Dimou, Y. Aribou, A. Chahbi and S. Kabbaj

Pm−1
Proof. Let D(x, y) = k=0 f (x + bk y) − mf (x). It is clear that
m−1
X m−1
X m−1
X
m[ (f (x + bk y) − mf (x)] = [ (mf (x + bk y) − f (x + bk y + bn t))]
k=0 k=0 n=0
m−1
X m−1
X
+ [ (−mf (x + bk t) + f (x + bk t + bn y))]
k=0 n=0
m−1
X
+ m[ f (x + bk t) − mf (x)],
k=0
m−1
X
= − D(x + bk y, t)
k=0
m−1
X
+ D(x + bk t, y) + mD(x, t).
k=0

Since Ω satisfies (C), for given x, y ∈ E, then there exists t ∈ E such that
kD(x + bk y, t)k ≤ δ , kD(x + bk t, y)k ≤ δ and kD(x, t)k ≤ δ. Now by using
the triangle inequality, we get
m−1
X
m|| (f (x + bk y) − mf (x)|| ≤ 3mδ, x, y ∈ E.
k=0

This implies that

m−1
X
|| (f (x + bk y) − mf (x)|| ≤ 3δ, x, y ∈ E.
k=0

Next, according to Theorem 1.1, there exists a unique generalized poly-

nomial (GP) q : E → F of degree at most m − 1 such that q(0) = 0
and
kf (x) − f (0) − q(x)k ≤ 3m δ f or all x ∈ E.
This completes the proof. 2
Now with m = 2. The condition (C) is reduced to the following: for
given x, y ∈ E there exists t ∈ E such that

(2.3) {(x + y, t), (x − y, t), (x + t, y), (x − t, y), (x, t)} ⊂ Ω.

As a direct consequence of Theorem 2.1, we obtain the following corol-
lary
 Stability of generalized Jensen functional equation on a set of ... 461

Corollary 2.2. Let δ ≥ 0. Suppose that f : E → F satisfies the functional

inequality

||f (x + y) + f (x − y) − 2f (x)|| ≤ δ f or all x, y ∈ Ω.

then there exists a unique additive mapping A : E → F such that

||f (x) − f (0) − A(x)|| ≤ 9δ f or all x ∈ E.

3. The Ulam-Hyers stability in a set of Lebesgue measure

zero

In this section, we prove the Ulam-Hyers stability in a set of Lebesgue

measure zero for the functional inequality
m−1
X
(3.1) || f (z + bk ξ) − mf (z)|| ≤ δ for all (z, ξ) ∈ Ω
k=0

where f : C → F and Ω ⊂ C2 is of four-dimensional Lebesgue measure

zero, and the Jensen functional inequality

(3.2) ||f (x + y) + f (x − y) − 2f (x)|| ≤ δ f or all (x, y) ∈ Ω ⊂ R2 ,

where f : R → F and Ω is of two-dimensional Lebesgue measure zero.

Firstly, the inequality (3.1) is a particular case of(1.3) where E = C.

Now, the condition (C) reduces to the following:

For given z, ξ ∈ C there exists η ∈ C such that

(3.3) {(z + bk ξ, η), (z + bk η, ξ), (z, η) : k = 0, . . . , m − 1} ⊂ Ω.

By virtue of Theorem 2.1, it suffice to construct a set Ω ⊂ C2 of measure
zero satisfying (3.3). It is known from [[21],Theorem 1.6] that there exists a
set K ⊂ R of Lebesgue measure 0 such that R \ K is of first Baire category.
That is, R \ K is a countable union of nowhere dense subsets of R.

Lemma 3.1 (Lemma 2.4, [9]). Let K be a subset of R of measure 0

such that K = R \ K is of first Baire category. Then, for any countable
subsets U ⊂ R,V ⊂ R\{0} and M > 0, there exists t ≥ M such that

(3.4) U + tV = {u + tv : u ∈ U, v ∈ V } ⊂ K.
462 H. Dimou, Y. Aribou, A. Chahbi and S. Kabbaj

From now on we identify C with R2 .

Theorem 3.2. Let K be the set defined in Lemma 3.1, R be the rotation
⎛ ⎞
− cos θ 0 sin θ 0
⎜ 0 − cos θ 0 sin θ ⎟
⎜ ⎟
⎜ ⎟
⎝ − sin θ 0 − cos θ 0 ⎠
0 − sin θ 0 − cos θ

with θ ∈ R\{ π2 + kπ, kπ/k ∈ Z} and Ω = R−1 (K × K × K × K). Then Ω

satisfies (3.3) and has four-dimensional Lebesgue measure 0.

Proof. Our method in this proof is inspired by [10, Theorem 3.2]. Let

z = x + iy, ξ = u + iv, η = t + is ∈ C, k = 0, 1, · · · , m − 1,

and let
Pz,ξ,η,k = {(x + u cos 2πk 2πk 2πk 2πk
m − v sin m , y + u sin m + v cos m , t, s)}
S
{(x + t cos 2πk 2πk 2πk 2πk
m − s sin m , y + t sin m + s cos m , u, v),

(x, y, t, s)}.

Then Ω satisfies (3.3) if and only if, for every z = x+iy and ξ = u+iv ∈
C, there exists η = t + is ∈ C such that
m−1
[
(3.5) R( Pz,ξ,η,k ) ⊂ K × K × K × K.
k=0

Hence the inclusion (3.5) is equivalent to

m−1
[
Sz,ξ,η = {− cos θp1 + sin θp3 , − cos θp2 + sin θp4 , − sin θp1 − cos θp3 ,
k=0
− sin θp2 − cos θp4 : (p1 , p2 , p3 , p4 ) ∈ Pz,ξ,η,k } ⊂ K.

Now, we can choose α ∈ R(α 6= 0) such that

2πk 2πk 2πk 2πk
cos − α sin 6= 0, sin + α cos 6= 0
m m m m
for all k = 0, 1, · · · , m − 1. Then it is easy to check that the set Sz,ξ,t+αti is
contained in the set of the form U + tV, where
 Stability of generalized Jensen functional equation on a set of ... 463

m−1
[ 2πk 2πk 2πk
U= {− cos θ(x + u cos − v sin ), − cos θ(y + u sin +
k=0
m m m

v cos 2πk 2πk 2πk 2πk

m ), − sin θ(x + u cos m − v sin m ), − sin θ(y + u sin m

+ v cos πk
m )}
S
{−x cos θ + u sin θ, −y cos θ + v sin θ, −(x sin θ + u cos θ), −(y sin θ

+v cos θ), −x cos θ, −y cos θ, −x sin θ, −y sin θ}

m−1
[ 2πk 2πk
V = {sin θ, α sin θ, − cos θ, −α cos θ, − cos θ(cos − α sin ),
k=0
m m

− cos θ(sin 2πk 2πk 2πk 2πk 2πk

m +α cos m ), − sin θ(cos m −α sin m ), − sin θ(sin m

+α cos 2πk
m )}.

By (3.4), for given z = x + iy, ξ = u + iv ∈ C and M > 0 there exists

t ≥ M such that
Sz,ξ,t+αti ⊂ U + tV ⊂ K.
Thus, Ω satisfies (3.3). 2

Theorem 3.3. There exists a set Ω ⊂ C2 of Lebesgue measure zero such

that if f : C → F satisfies the inequality
m−1
X
|| f (z + bk ξ) − mf (z)|| ≤ δ for all(z, ξ) ∈ Ω
k=0
for all (z, ξ) ∈ Ω, then there exists a unique mapping q : C → F satisfying
m−1
X
q(z + bk ξ) − mq(z) = 0 for all z, ξ ∈ C.
k=0
such that

||f (z) − f (0) − q(z)|| ≤ 3m δ f or all z ∈ C.

464 H. Dimou, Y. Aribou, A. Chahbi and S. Kabbaj

Proof. The proof follows immediately from Theorems 3.2 and 2.1. 2

Remark 3.4. It is easy to see that the set Ωd = {(z, ξ) ∈ C2 : |z|+|ξ| ≥ d}

also satisfies (3.3). Thus, the result of Theorem (3.3) holds when Ω is
replaced by Ωd . Thus, as consequence of Theorem (3.3) with the above
remark, we obtain a strong version of asymptotic behavior of f satisfies

m−1
X
(3.6) || f (z + bk ξ) − mf (z)|| → 0, as |z| + |ξ| → ∞, (z, ξ) ∈ Ω.
k=0

Corollary 3.5. Suppose that f : C → F satisfies (3.6). Then f satisfies

the functional equation
m−1
X
f (z + bk ξ) = mf (z), z, ξ ∈ C.
k=0

Secondly, we consider (3.2) in view of Corollary 2.2, it suffice to con-

struct a set Ω ⊂ R2 of measure zero satisfying (2.3).

Theorem 3.6. Let Ω = exp(−iθ)K × K be the rotation of K × K by −θ.

Then Ω satisfies (2.3) and has two-dimensional Lebesgue measure 0.

Proof. The proof is similar to the proof of Theorem (3.2). However, we

prove the completeness. Let x, y, t ∈ R, and let

Px,y,t = {(x + y, t), (x − y, t), (x + t, y), (x − t, y), (x, t)}.

Then Ω satisfies (2.3) if and only if, for every x, y ∈ R such that

(3.7) exp(iθ)Px,y,t ⊂ K × K.

The inclusion (3.7) is equivalent to

Sx,y,t := {− cos θu + sin θv, − sin θu − cos θv : (u, v) ∈ Px,y,t } ⊂ K.

It is easy to check that the set Sx,y,t is contained in a set of form U +tV ,
where
U = {cosθ(x + y), sinθ(x + y), cosθ(x − y), sin θ(x − y), x cos θ
−y sin θ, y cos θ + x sin θ, x cos θ, x sin θ}
V = {± sinθ, ; ± cosθ}
 Stability of generalized Jensen functional equation on a set of ... 465

By Lemma 3.1, for given x, y ∈ R and M > 0 there exists t ≥ M such

that
Sx,y,t = U + tV ⊂ K.
Thus, Ω satisfies (2.3). 2

By Corollary 2.2 and Theorem (3.6) we have the following result.

Theorem 3.7. Let δ ≥ 0. There exists a set Ω ⊂ R2 of Lebesgue measure

zero such that if f : R → F satisfies the inequality

||f (x + y) + f (x − y) − 2f (x)|| ≤ δ.

only for all x, y ∈ Ω, then there exists a unique additive mapping A : R → F

such that
||f (x) − f (0) − A(x)|| ≤ 9δ f or all x ∈ R.

: The authors are very thankful to the referee for valuable suggestions.

References

[1] T. Aoki, On the stability of the linear transformation in Banach spaces,

Journal of the Mathematical Society of Japan, 2, pp. 64-66, (1950).

[2] J. A. Baker, A general functional equation and its stability, Proceeding

of the American Mathematical Society, 133, pp. 1657-1664, (2005).

[3] N. Brillouet-Belluot, J. Brzdȩk, K. Ciepliǹski, On some recent develop-

ments in Ulam’s type stability, Abstract and Applied Analysis, (2012).

[4] J. Brzdȩk, On a method of proving the Hyers-Ulam stability of func-

tional equations on restricted domains, The Australian Journal of
Mathematical Analysis and Applications, 6, pp. 1-10, (2009).

[5] A. B. Chahbi, A. Charifi, B. Bouikhalene and S. Kabbaj, Non-

archimedean stability of a Pexider K-quadratic functional equation,
Arab Journal of Mathematical Sciences, 21, pp. 67-83, (2015).
466 H. Dimou, Y. Aribou, A. Chahbi and S. Kabbaj

[6] A. Chahbi, M. Almahalebi, A. Charifi and S.Kabbaj Generalized

Jensen functional equation on restricted domain, Annals of West Uni-
versity of Timisoara-Mathematics, 52, pp. 29-39, (2014).

[7] P. W. Cholewa, Remarks on the stability of functional equations, Ae-

quationes mathematicae, 27, pp. 76-86, (1984).

[8] J. Chung , Stability of a conditional Cauchy equation on a set of mea-

sure zero, Aequationes mathematicae, 87, pp. 391-400, (2014).

[9] J. Chung and J. M. Rassias, Quadratic functional equations in a set

of Lebesgue measure zero, Journal of Mathematical Analysis and Ap-
plications, 419, pp. 1065-1075, (2014).

[10] J. Chung and J. M. Rassias, On a measure zero Stability problem of a

cyclic equation, Bulletin of the Australian Mathematical Society, 93,
pp. 1-11, (2016).

[11] S. Czerwik, On the stability of the quadratic mapping in normed

spaces, Abhandlungen aus dem Mathematischen Seminar der Univer-
sität Hamburg, 62, pp. 59-64, (1992).

[12] P. Gǎvruta, A generalization of the Hyers-Ulam-Rassias stability of

approximately additive mappings, Journal of Mathematical Analysis
and Applications, 184, pp. 431-436, (1994).

[13] D. H. Hyers, On the stability of the linear functional equation, Proceed-

ings of the National Academy of Sciences, 27, pp. 222-224, (1941).

[14] D. H. Hyers and Th. M. Rassias, Approximate homomorphisms, Ae-

quationes Mathematicae, 44, pp. 125-153, (1992).

[15] D. H. Hyers, Transformations with bounded n-th differences, Pacific

Journal of Mathematics, 11, pp. 591-602, (1961).

[16] K. W. Jun and Y. H. Lee, A generalization of the Hyers-Ulam-Rassias

stability of Jensen0 s equation, Journal of Mathematical Analysis and
Applications, 238, pp. 305-315, (1999).

[17] S. M. Jung, On the HyersUlam stability of the functional equations

that have the quadratic property, Journal of Mathematical Analysis
and Applications, 222, pp. 126-137, (1998).
 Stability of generalized Jensen functional equation on a set of ... 467

[18] C. F. K. Jung, On generalized complete metric spaces, Bulletin of the

American Mathematical Society, 75, pp. 113-116, (1969).

[19] R. LÃ ukasik, Some generalization of Cauchy0 s and the quadratic func-

tional equations, Aequationes Mathematicae, 83, pp. 75-86, (2012).

[20] A. Najati, S. M. Jung, Approximately quadratic mappings on restricted

domains, Journal of Inequalities and Applications, (2010).

[21] J. C. Oxtoby, Measure and Category, Springer, New-York (1980).

[22] Th. M. Rassias, On the stability of linear mapping in Banach spaces,

Proceedings of the American Mathematical Society, 72, pp. 297-300,
(1978).

[23] Th. M. Rassias, On the stability of the functional equations and a

problem of Ulam, Acta Applicandae Mathematicae, 62, pp. 23-130,
(2000).

[24] Th. M. Rassias and P. Šemrl, On the behavior of mappings which do

not satisfy Hyers-Ulam stability, Proceedings of the American Mathe-
matical Society, 114, pp. 989-993, (1992).

[25] Th. M. Rassias, and J. Tabor, Stability of Mappings of Hyers-Ulam

Type, Hardronic Press, (1994).

[26] J. M. Rassias, On the Ulam stability of mixed type mappings on re-

stricted domains, Journal of Mathematical Analysis and Applications,
276, pp. 747-762, (2002).

[27] F. Skof, Local properties and approximations of operators, Rendiconti

del Seminario Matematico e Fisico di Milano, 53, pp. 113-129, (1983).

[28] H. Stetkær, Functional equations involving means of functions on the

complex plane, Aequationes Mathematicae, 56, pp. 47-62, (1998).

[29] S. M. Ulam, A Collection of Mathematical Problems, Interscience Pub-

lishers, 8(1960).
468 H. Dimou, Y. Aribou, A. Chahbi and S. Kabbaj

Hajira Dimou
Department of Mathematics,
Faculty of Sciences,
IBN Tofail University,
BP: 14000, Kenitra,
Morocco
e-mail : [email protected]

Youssef Aribou
Department of Mathematics,
Faculty of Sciences,
IBN Tofail University,
BP: 14000, Kenitra,
Morocco
e-mail : [email protected]

Abdellatif Chahbi
Department of Mathematics,
Faculty of Sciences,
IBN Tofail University,
BP: 14000, Kenitra,
Morocco
e-mail : [email protected]

and

Samir Kabbaj
Department of Mathematics,
Faculty of Sciences,
IBN Tofail University,
BP: 14000, Kenitra,
Morocco
e-mail : [email protected]