Next Article in Journal
Fractional Calculus as a Simple Tool for Modeling and Analysis of Long Memory Process in Industry
Previous Article in Journal
On History of Mathematical Economics: Application of Fractional Calculus
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

On the Hyers-Ulam-Rassias Stability of a General Quintic Functional Equation and a General Sextic Functional Equation

Department of Mathematics Education, Gongju National University of Education, Gongju 32553, Korea
Mathematics 2019, 7(6), 510; https://doi.org/10.3390/math7060510
Submission received: 14 May 2019 / Revised: 30 May 2019 / Accepted: 31 May 2019 / Published: 4 June 2019

Abstract

:
The general quintic functional equation and the general sextic functional equation are generalizations of many functional equations such as the additive function equation and the quadratic function equation. In this paper, we investigate Hyers–Ulam–Rassias stability of the general quintic functional equation and the general sextic functional equation.

1. Introduction

Let X be a real normed space and Y be a real Banach space. In 1940, Ulam [1] raised the question about the stability of group of homomorphisms, and in the following year, Hyers [2] solved this question about the additive functional equation, which gave a partial answer to Ulam’s question. In 1978, Rassias [3] generalized Hyers’ result (refer to [4,5,6,7,8] for a more generalized result). Since then, many mathematicians have investigated the stability of different types of functional equations [9,10]. Rassias [3] investigated the stability problem for approximately linear mappings controlled by the unbounded function θ ( x p + y p ) as follow:
Theorem 1.
Let f : X Y be a mapping from a real normed vector space X into a Banach space Y satisfying the inequality:
f ( x + y ) f ( x ) f ( y ) θ ( x p + y p ) ,
for all x , y X \ { 0 } , where θ and p are constants with θ > 0 and p < 1 . If f ( t x ) is continuous in t for each fixed x, then there exists a unique linear mapping T : X Y such that:
f ( x ) T ( x ) 2 θ x p | 2 2 p | ,
for all x X \ { 0 } .
The functional equation is said to have Hyers–Ulam–Rassias stability when the stability can be proven under the control function θ ( x p + y p ) .
A mapping f : X Y is called a general quintic mapping if f satisfies the functional equation:
i = 0 6 6 C i ( 1 ) 6 i f ( x + ( i 3 y ) ) = 0
which is called a general quintic functional equation. A mapping f : X Y is called a general sextic mapping:
i = 0 7 7 C i ( 1 ) 7 i f ( x + i y ) = 0
which is called a general sextic functional equation. For example, the functions f , g : R R , defined by f ( x ) = i = 0 5 a i x i and g ( x ) = i = 0 6 a i x i , a i R , satisfy the above functional equations. More detailed terms for the concepts of “a general quintic mapping” and “a general sextic mapping” can be found in Baker’s paper [11] by the terms “generalized polynomial mapping of degree at most 5” and “generalized polynomial mapping of degree at most 6”, respectively. Kim et al. [12] previously studied the stability of a general a general cubic functional equation, and Lee [13,14,15] studied the stability of a general quadratic functional equation, a general cubic functional equation, and a general quartic functional equation.
In Section 2, we will investigate the Hyers–Ulam–Rassias stability of the general quintic functional equation. In Section 3, we will investigate the Hyers–Ulam–Rassias stability of the general sextic functional equation.

2. Stability of a General Quintic Functional Equation

Throughout this section, for a given mapping f : X Y , we use the following abbreviations:
f o ( x ) : = f ( x ) f ( x ) 2 , f e ( x ) : = f ( x ) + f ( x ) 2 , D f ( x , y ) : = i = 0 6 6 C i ( 1 ) 6 i f ( x + ( i 3 y ) ) , Γ f ( x ) : = D f o ( 2 x , 2 x ) + 6 D f o ( 3 x , x ) + 36 D f o ( 2 x , x ) + 70 D f o ( x , x ) , Δ f ( x ) : = D f e ( x , x ) + 3 D f e ( 0 , x )
for all x , y X . By laborious computation, we can get the equalities:
Γ f ( x ) = f o ( 8 x ) 42 f o ( 4 x ) + 336 f o ( 2 x ) 512 f o ( x ) ,
Δ f ( x ) = f e ( 4 x ) 20 f e ( 2 x ) + 64 f e ( x )
for all x X .
Lemma 1.
Let p be a fixed nonnegative real number such that p { 1 , 2 , 3 , 4 , 5 } . For a given mapping f : X Y with f ( 0 ) = 0 , let J n f : X Y be the mappings defined by:
J n f ( x ) : = 4 n 3 f e x 2 n 16 f e x 2 n + 1 + 16 n + 1 12 f e x 2 n 4 f e x 2 n + 1 + 2 n 20 × 8 n + 64 × 32 n 45 f o x 2 n 40 × 2 n 680 × 8 n + 640 × 32 n 45 f o x 2 n + 1 + 256 × 2 n 1280 × 8 n + 1024 × 32 n 45 f o x 2 n + 2 i f 5 < p , 4 n 3 f e x 2 n 16 f e x 2 n + 1 + 16 n + 1 12 f e x 2 n 4 f e x 2 n + 1 + 2 n 5 × 8 n 90 f o x 2 n 1 40 × 2 n 170 × 8 n 90 f o x 2 n + 256 × 2 n 320 × 8 n 90 f o x 2 n + 1 + 4 90 × 32 n f o ( 2 n + 1 x ) 10 f o ( 2 n x ) + 16 f o ( 2 n 1 x ) i f 4 < p < 5 , 4 n 12 16 f e ( 2 n x ) f e ( 2 n + 1 x ) 4 f e ( 2 n x ) f e ( 2 n + 1 x ) 12 × 16 n + 2 n 5 × 8 n 90 f o x 2 n 1 40 × 2 n 170 × 8 n 90 f o x 2 n + 256 × 2 n 320 × 8 n 90 f o x 2 n + 1 + 4 90 × 32 n f o ( 2 n + 1 x ) 10 f o ( 2 n x ) + 16 f o ( 2 n 1 x ) i f 3 < p < 4 , 4 n 12 16 f e ( 2 n x ) f e ( 2 n + 1 x ) 4 f e ( 2 n x ) f e ( 2 n + 1 x ) 12 × 16 n + 4 f o ( 2 n + 1 x ) 90 × 32 n 40 f o ( 2 n x ) 90 × 32 n + 64 f o ( 2 n 1 x ) 90 × 32 n 5 f o ( 2 n + 1 x ) 90 × 8 n + 170 f o ( 2 n x ) 90 × 8 n 320 f o ( 2 n 1 x ) 90 × 8 n + 2 n 90 f o x 2 n 1 40 × 2 n 90 f o x 2 n + 256 × 2 n 90 f o x 2 n + 1 i f 2 < p < 3 , 16 f e ( 2 n x ) f e ( 2 n + 1 x ) 12 × 4 n 4 f e ( 2 n x ) f e ( 2 n + 1 x ) 12 × 16 n + 4 f o ( 2 n + 1 x ) 90 × 32 n 40 f o ( 2 n x ) 90 × 32 n + 64 f o ( 2 n 1 x ) 90 × 32 n 5 f o ( 2 n + 1 x ) 90 × 8 n + 170 f o ( 2 n x ) 90 × 8 n 320 f o ( 2 n 1 x ) 90 × 8 n + 2 n 90 f o x 2 n 1 40 × 2 n 90 f o x 2 n + 256 × 2 n 90 f o x 2 n + 1 i f 1 < p < 2 , 16 f e ( 2 n x ) f e ( 2 n + 1 x ) 12 × 4 n 4 f e ( 2 n x ) f e ( 2 n + 1 x ) 12 × 16 n + f o ( 2 n + 2 x ) 720 × 32 n 10 f o ( 2 n + 1 x ) 720 × 32 n + 16 f o ( 2 n x ) 720 × 32 n 5 f o ( 2 n + 2 x ) 720 × 8 n + 170 f o ( 2 n + 1 x ) 720 × 8 n 320 f o ( 2 n x ) 720 × 8 n + f o ( 2 n + 2 x ) 40 f o ( 2 n + 1 x ) + 256 f o ( 2 n x ) 180 × 2 n i f 0 p < 1
for all x X and all nonnegative integers n. Then,
J n f ( x ) J n + 1 f ( x ) = :         4 2 n + 1 3 4 n 3 Δ f x 2 n + 2 + 2 n 45 4 × 8 n 9 + 64 × 32 n 45 Γ f x 2 n + 3 i f 5 < p , 4 2 n + 1 3 4 n 3 Δ f x 2 n + 2 + 2 n 90 8 n 18 Γ f x 2 n + 2 2 Γ f ( 2 n 1 x ) 45 × 32 n + 1 i f 4 < p < 5 , 4 n 12 Δ f x 2 n + 1 Δ f ( 2 n x ) 192 × 16 n + 2 n 90 8 n 18 Γ f x 2 n + 2 2 Γ f ( 2 n 1 x ) 45 × 32 n + 1 i f 3 < p < 4 , 4 n 12 Δ f x 2 n + 1 Δ f ( 2 n x ) 192 × 16 n + 2 n 90 Γ f x 2 n + 2 + Γ f ( 2 n 1 x ) 18 × 8 n + 1 2 Γ f ( 2 n 1 x ) 45 × 32 n + 1 i f 2 < p < 3 , Δ f ( 2 n x ) 48 × 4 n Δ f ( 2 n x ) 192 × 16 n + 2 n 90 Γ f x 2 n + 2 + Γ f ( 2 n 1 x ) 18 × 8 n + 1 2 Γ f ( 2 n 1 x ) 45 × 32 n + 1 i f 1 < p < 2 , Δ f ( 2 n x ) 48 × 4 n Δ f ( 2 n x ) 192 × 16 n + Γ f ( 2 n x ) 180 × 2 n + 1 + Γ f ( 2 n x ) 144 × 8 n + 1 Γ f ( 2 n x ) 720 × 32 n + 1 i f 0 p < 1
for all x X and all nonnegative integers n.
Proof. 
For the case 2 < p < 3 , from the definition of J n f and the equalities (3), we obtain that:
J n f ( x ) J n + 1 f ( x )     = 4 n 12 16 f e x 2 n f e x 2 n 1 4 n + 1 12 16 f e x 2 n + 1 f e x 2 n         4 f e ( 2 n x ) f e ( 2 n + 1 x ) 12 × 16 n + 4 f e ( 2 n + 1 x ) f e ( 2 n + 2 x ) 12 × 16 n + 1         + 4 f o ( 2 n + 1 x ) 90 × 32 n 40 f o ( 2 n x ) 90 × 32 n + 64 f o ( 2 n 1 x ) 90 × 32 n         4 f o ( 2 n + 2 x ) 90 × 32 n + 1 + 40 f o ( 2 n + 1 x ) 90 × 32 n + 1 64 f o ( 2 n x ) 90 × 32 n + 1         5 f o ( 2 n + 1 x ) 90 × 8 n + 170 f o ( 2 n x ) 90 × 8 n 320 f o ( 2 n 1 x ) 90 × 8 n         + 5 f o ( 2 n + 2 x ) 90 × 8 n + 1 170 f o ( 2 n + 1 x ) 90 × 8 n + 1 + 320 f o ( 2 n x ) 90 × 8 n + 1         + 2 n 90 f o x 2 n 1 40 f o x 2 n + 256 f o x 2 n + 1         2 n + 1 90 f o x 2 n 40 f o x 2 n + 1 + 256 f o x 2 n + 2     = 4 n 12 f e x 2 n 1 + 20 f e x 2 n 64 f e x 2 n + 1         f e ( 2 n + 2 x ) 20 f e ( 2 n + 1 x ) + 64 f e ( 2 n x ) 12 × 16 n + 1         4 f o ( 2 n + 2 x ) 168 f o ( 2 n + 1 x ) + 1344 f o ( 2 n x ) 2048 f o ( 2 n 1 x ) 90 × 32 n + 1         + 5 f o ( 2 n + 2 x ) 210 f o ( 2 n + 1 x ) + 1680 f o ( 2 n x ) 2560 f o ( 2 n 1 x ) 90 × 8 n + 1         + 2 n 90 f o x 2 n 1 42 f o x 2 n + 336 f o x 2 n + 1 512 f o x 2 n + 2     = 4 n 12 Δ f x 2 n + 1 Δ f ( 2 n x ) 192 × 16 n + 2 n 90 Γ f x 2 n + 2 + Γ f ( 2 n 1 x ) 18 × 8 n + 1 2 Γ f ( 2 n 1 x ) 45 × 32 n + 1
for all x X and all nonnegative integers n. Furthermore, we easily show that the equality (5) holds by a similar method for the other cases, either 0 p < 1 , or 1 < p < 2 , or 3 < p < 4 , or 4 < p < 5 , or 5 < p . □
Lemma 2.
If f : X Y is a mapping such that:
D f ( x , y ) = 0
for all x , y X with f ( 0 ) = 0 , then
J n f ( x ) = f ( x )
for all x X and all positive integers n.
Proof. 
If f : X Y is a mapping such that:
D f ( x , y ) = 0
for all x , y X with f ( 0 ) = 0 , then it follows from the definitions of Δ f ( x ) and Γ f ( x ) that Δ f ( x ) = 0 and Γ f ( x ) = 0 for all x X . Therefore, together with the equality f ( x ) J n f ( x ) = i = 0 n 1 ( J i f ( x ) J i + 1 f ( x ) ) and the equality (6), we conclude that:
J n f ( x ) = f ( x )
for all x X and all positive integers n. □
From Lemma 2, we can prove the following stability theorem.
Theorem 2.
Let p 1 , 2 , 3 , 4 , 5 be a fixed nonnegative real number. Suppose that f : X Y is a mapping such that:
D f ( x , y ) θ ( x p + y p )
for all x , y X . Then, there exists a general quintic mapping F such that
f ( x ) f ( 0 ) F ( x ) :         K 45 × 2 2 p ( 2 p 2 ) + ( 128 + 44 × 2 p ) K 45 ( 2 p 32 ) ( 2 p 8 ) 2 2 p + 5 2 p ( 2 p 16 ) ( 2 p 4 ) θ x p i f 5 < p , ( 2 × 2 p 1 ) K 45 ( 2 p 8 ) ( 2 p 2 ) 2 p + 2 K 45 ( 32 2 p ) 2 p + 5 2 p ( 2 p 16 ) ( 2 p 4 ) θ x p i f 4 < p < 5 , ( 2 × 2 p 1 ) K 45 ( 2 p 8 ) ( 2 p 2 ) 2 p + 2 K 45 ( 32 2 p ) 2 p + 5 ( 2 p 4 ) ( 16 2 p ) θ x p i f 3 < p < 4 , K 90 × 2 p ( 2 p 2 ) + ( 128 2 p ) K 90 ( 32 2 p ) ( 8 2 p ) 2 p + 5 ( 2 p 4 ) ( 16 2 p ) θ x p i f 2 < p < 3 , K 90 × 2 p ( 2 p 2 ) + ( 128 2 p ) K 90 ( 32 2 p ) ( 8 2 p ) 2 p + 5 ( 16 2 p ) ( 4 2 p ) θ x p i f 1 < p < 2 , K 180 ( 2 2 p ) + ( 38 2 p ) K 180 ( 8 2 p ) ( 32 2 p ) + 5 ( 16 2 p ) ( 4 2 p ) θ x p i f 0 p < 1
for all x X and F ( 0 ) = 0 , where K = 182 + 38 × 2 p + 6 × 3 p .
Proof. 
If f ˜ is the mapping defined by f ˜ ( x ) = f ( x ) f ( 0 ) , then the mapping f ˜ satisfies the properties D f ˜ ( x , y ) = D f ( x , y ) and f ˜ ( 0 ) = 0 . By (6) and the definitions of Γ f and Δ f , we have:
Γ f ˜ ( x ) = D f o ( 2 x , 2 x ) + 6 D f o ( 3 x , x ) + 36 D f o ( 2 x , x ) + 70 D f o ( x , x ) θ ( 182 + 38 × 2 p + 6 × 3 p ) x p , Δ f ˜ ( x ) = D f e ( x , x ) + 3 D f e ( 0 , x ) 5 θ x p
for all x X . It follows from (5) and (6) that
J n f ˜ ( x ) J n + 1 f ˜ ( x ) :         2 n K 45 × 2 ( n + 3 ) p 4 × 8 n K 9 × 2 ( n + 3 ) p + 64 × 32 n K 45 × 2 ( n + 3 ) p + 5 ( 4 2 n + 1 4 n ) 3 × 2 ( n + 2 ) p θ x p if 5 < p , ( 5 × 8 n 2 n ) K 90 × 2 ( n + 2 ) p + 2 × 2 ( n 1 ) p K ) 45 × 32 n + 1 + 5 ( 4 2 n + 1 4 n ) 3 × 2 ( n + 2 ) p θ x p if 4 < p < 5 , ( 5 × 8 n 2 n ) K 90 × 2 ( n + 2 ) p + 2 × 2 ( n 1 ) p K ) 45 × 32 n + 1 + 4 n 12 5 2 ( n + 1 ) p + 5 × 2 n p 192 × 16 n θ x p if 3 < p < 4 , 2 n K 90 × 2 ( n + 2 ) p + ( 5 × 4 n + 1 4 ) × 2 ( n 1 ) p K 90 × 32 n + 1 + 4 n 12 5 2 ( n + 1 ) p + 5 × 2 n p 192 × 16 n θ x p if 2 < p < 3 , 2 n K 90 × 2 ( n + 2 ) p + ( 5 × 4 n + 1 4 ) × 2 ( n 1 ) p K 90 × 32 n + 1 + 5 ( 4 n + 1 1 ) 2 n p 192 × 16 n θ x p if 1 < p < 2 , 2 n p K 180 × 2 n + 1 + 2 n p K 144 × 8 n + 1 2 n p K 720 × 32 n + 1 + 5 ( 4 n + 1 1 ) 2 n p 192 × 16 n θ x p if 0 p < 1
for all x X . Together with the equality J n f ˜ ( x ) J n + m f ˜ ( x ) = i = n n + m 1 ( J i f ˜ ( x ) J i + 1 f ˜ ( x ) ) for all x X , we obtain that
J n f ˜ ( x ) J n + m f ˜ ( x ) :         i = n n + m 1 ( 2 i 20 × 8 i + 64 × 32 i ) K 45 × 2 ( i + 3 ) p + 5 ( 4 2 i + 1 4 i ) 3 × 2 ( i + 2 ) p θ x p if 5 < p , i = n n + m 1 ( 5 × 8 i 2 i ) K 90 × 2 ( i + 2 ) p + 2 × 2 ( i 1 ) p K ) 45 × 32 i + 1 + 5 ( 4 2 i + 1 4 i ) 3 × 2 ( i + 2 ) p θ x p if 4 < p < 5 , i = n n + m 1 ( 5 × 8 i 2 i ) K 90 × 2 ( i + 2 ) p + 2 × 2 ( i 1 ) p K ) 45 × 32 i + 1 + 4 i 12 5 2 ( i + 1 ) p + 5 × 2 i p 192 × 16 i θ x p if 3 < p < 4 , i = n n + m 1 32 × 64 i K + ( 5 × 4 i + 1 4 ) × 2 ( 2 i + 1 ) p K 90 × 32 i + 1 × 2 ( i + 2 ) p + 4 i 12 5 2 ( i + 1 ) p + 5 × 2 i p 192 × 16 i θ x p if 2 < p < 3 , i = n n + m 1 32 × 64 i K + ( 5 × 4 i + 1 4 ) × 2 ( 2 i + 1 ) p K 90 × 32 i + 1 × 2 ( i + 2 ) p + 5 ( 4 i + 1 1 ) 2 i p 192 × 16 i θ x p if 1 < p < 2 , i = n n + m 1 2 i p K 180 × 2 i + 1 + 2 i p K 144 × 8 i + 1 2 i p K 720 × 32 i + 1 + 5 ( 4 i + 1 1 ) 2 i p 192 × 16 i θ x p if 0 p < 1
for all x X and n , m N { 0 } . It follows from (8) that the sequence { J n f ˜ ( x ) } is a Cauchy sequence for all x X . Since Y is complete, the sequence { J n f ˜ ( x ) } converges for all x X . Hence, we can define a mapping F : X Y by:
F ( x ) : = lim n J n f ˜ ( x )
for all x X . Note that F ( 0 ) = 0 follows from f ˜ ( 0 ) = 0 . Moreover, letting n = 0 and passing the limit n in (8), we get the inequality (7). For the case 2 < p < 3 , from the definition of F, we easily get:
D F ( x , y ) = lim n 4 n 12 D f e 2 x 2 n , 2 y 2 n + 16 D f e x 2 n , y 2 n         + D f e 2 n + 1 x , 2 n + 1 y 4 D f e 2 n x , 2 n y 12 × 16 n         + 4 D f o ( 2 n + 1 x , 2 n + 1 y ) 90 × 32 n 40 D f o ( 2 n x , 2 n y ) 90 × 32 n + 64 D f o ( 2 n 1 x , 2 n 1 y ) 90 × 32 n         5 [ D f o ( 2 n + 1 x , 2 n + 1 y ) 34 D f o ( 2 n x , 2 n y ) + 64 D f o ( 2 n 1 x , 2 n 1 y ) ] 90 × 8 n         + 2 n 90 D f o 2 x 2 n , 2 y 2 n 40 D f o x 2 n , y 2 n + 256 D f o x 2 n + 1 , y 2 n + 1 lim n ( 4 n ( 2 p + 16 ) 12 × 2 n p + 2 n p ( 2 p + 4 ) 12 × 16 n + 4 ( 4 p + 10 × 2 p + 16 ) 2 ( n 1 ) p 90 × 32 n         + 5 ( 4 p + 34 × 2 p + 64 ) 2 ( n 1 ) p 90 × 8 n + 2 n ( 4 p + 40 × 2 p + 256 ) 90 × 2 ( n + 1 ) p ) × θ ( x p + y p ) = 0
for all x , y X . Furthermore, we easily show that D F ( x , y ) = 0 by a similar method for the other cases, either 0 p < 1 , or 1 < p < 2 , or 3 < p < 4 , or 4 < p < 5 , or 5 < p . To prove the uniqueness of F, let F : X Y be another general quintic mapping satisfying (7) and F ( 0 ) = 0 . By Lemma 2, the equality F ( x ) = J n F ( x ) holds for all n N . For the case 2 < p < 3 , we have:
J n f ˜ ( x ) F ( x ) = J n f ˜ ( x ) J n F ( x ) 4 n + 2 12 ( f ˜ e F e ) x 2 n + 4 n 12 ( f ˜ e F e ) 2 x 2 n + 4 ( f ˜ e F e ) ( 2 n x ) 12 × 16 n         + ( f ˜ e F e ) ( 2 n + 1 x ) 12 × 16 n + 4 ( f ˜ o F o ) ( 2 n + 1 x ) 90 × 32 n + 40 ( f ˜ o F o ) ( 2 n x ) 90 × 32 n         + 64 ( f ˜ o F o ) ( 2 n 1 x ) 90 × 32 n + 5 ( f ˜ o F o ) ( 2 n + 1 x ) 90 × 8 n + 170 ( f ˜ o F o ) ( 2 n x ) 90 × 8 n         + 320 ( f ˜ o F o ) ( 2 n 1 x ) 90 × 8 n + 2 n 90 ( f ˜ o F o ) 2 x 2 n         + 40 × 2 n 90 ( f ˜ o F o ) x 2 n + 256 × 2 n 90 ( f ˜ o F o ) x 2 n + 1 ( 4 n + 2 + 4 n 2 p 12 × 2 n p + 4 × 2 n p + 2 ( n + 1 ) p 12 × 16 n + 4 ( 4 p + 10 × 2 p + 16 ) 2 ( n 1 ) p 90 × 32 n         + 5 ( 4 p + 34 × 2 p + 64 ) 2 ( n 1 ) p 90 × 8 n + ( 4 p + 40 × 2 p + 256 ) 2 n 90 × 2 ( n + 1 ) p )         × K 90 × 2 p ( 2 p 2 ) + ( 128 2 p ) K 90 ( 32 2 p ) ( 8 2 p ) 2 p + 5 ( 2 p 4 ) ( 16 2 p ) θ x p
for all x X and all positive integers n. Taking the limit in the above inequality as n , we obtain the equality F ( x ) = lim n J n f ˜ ( x ) for all x X , which means that F ( x ) = F ( x ) for all x X . Furthermore, we easily show that F ( x ) = F ( x ) by a similar method for the other cases, either 0 p < 1 , or 1 < p < 2 , or 3 < p < 4 , or 4 < p < 5 , or 5 < p . □
When n is a fixed number such that n { 1 , 2 , 3 , 4 , 5 } , if f : R R is a solution of the functional equation i = 0 n n C i ( 1 ) i f ( i x + y ) n ! f ( x ) = 0 for all x , y R , then f : R R is a solution of the functional equation D f ( x , y ) = 0 for all x , y R .
Therefore, Example 1 in [16] shows that the assumption p 1 , 2 , 3 , 4 , 5 cannot be omitted in Theorem 2.
Example 1.
(Example 1 in [16])There is a mapping f : R R :
i = 0 n n C i ( 1 ) i f ( i x + y ) n ! f ( x ) 4 × ( n + 1 ) ! ( n + 1 ) 2 n ( | x | n + | y | n ) .
for all x , y R , but there do not exist a mapping F : R R and a constant d > 0 such that i = 0 n n C i ( 1 ) i F ( i x + y ) n ! F ( x ) = 0 and | f ( x ) F ( x ) | d | x | n for all x R .

3. Stability of a General Sextic Functional Equation

Throughout this section, for a given mapping f : X Y , we use the following abbreviations:
D f ( x , y ) : = i = 0 7 7 C i ( 1 ) 7 i f ( x + i y ) , Γ ( x ) : = D f o ( 6 x , 2 x ) + 6 D f o ( x , x ) + 42 D f o ( 2 x , x ) + 112 D f o ( 3 x , x ) , Δ f ( x ) : = D f e ( 6 x , 2 x ) + 8 D f e ( x , x ) + 56 D f e ( 2 x , x ) + 112 D f e ( 3 x , x )
for all x , y X . By laborious computation, we can get the equalities:
Γ f ( x ) = f o ( 8 x ) 42 f o ( 4 x ) + 336 f o ( 2 x ) 512 f o ( x ) ,
Δ f ( x ) = f e ( 8 x ) 84 f e ( 4 x ) + 1344 f e ( 2 x ) 4096 f e ( x )
for all x X .
The proofs of the following two lemmas are very similar to the proofs of Lemmas 1 and 2, so we omit them and just describe them.
Lemma 3.
Let p 1 , 2 , 3 , 4 , 5 , 6 be a fixed real number. For a given mapping f : X Y with f ( 0 ) = 0 , let J n f : X Y be the mappings defined by:
J n f ( x ) : = 4 n 20 × 16 n + 64 × 64 n 45 f e x 2 n 80 × 4 n 1360 × 16 n + 1280 × 64 n 45 f e x 2 n + 1 + 1024 × 4 n 5120 × 16 n + 4096 × 64 n 45 f e x 2 n + 2 + 2 n 20 × 8 n + 64 × 32 n 45 f o x 2 n 40 × 2 n 680 × 8 n + 640 × 32 n 45 f o x 2 n + 1 + 256 × 2 n 1280 × 8 n + 1024 × 32 n 45 f o x 2 n + 2 i f 6 < p , 4 n 5 × 16 n 180 f e x 2 n 1 80 × 4 n 340 × 16 n 180 f e x 2 n + 1024 × 4 n 1280 × 16 n 180 f e x 2 n + 1 + 4 180 × 64 n f e ( 2 n + 1 x ) 20 f e ( 2 n x ) + 64 f e ( 2 n 1 x ) + 2 n 20 × 8 n + 64 × 32 n 45 f o x 2 n 40 × 2 n 680 × 8 n + 640 × 32 n 45 f o x 2 n + 1 + 256 × 2 n 1280 × 8 n + 1024 × 32 n 45 f o x 2 n + 2 i f 5 < p < 6 , 4 n 5 × 16 n 180 f e x 2 n 1 80 × 4 n 340 × 16 n 180 f e x 2 n + 1024 × 4 n 1280 × 16 n 180 f e x 2 n + 1 + 4 180 × 64 n f e ( 2 n + 1 x ) 20 f e ( 2 n x ) + 64 f e ( 2 n 1 x ) + 2 n 5 × 8 n 90 f o x 2 n 1 40 × 2 n 170 × 8 n 90 f o x 2 n + 256 × 2 n 320 × 8 n 90 f o x 2 n + 1 + 4 90 × 32 n f o ( 2 n + 1 x ) 10 f o ( 2 n x ) + 16 f o ( 2 n 1 x ) i f 4 < p < 5 , 4 f e ( 2 n + 1 x ) 180 × 64 n 80 f e ( 2 n x ) 180 × 64 n + 256 f e ( 2 n 1 x ) 180 × 64 n 5 f e ( 2 n + 1 x ) 180 × 16 n + 340 f e ( 2 n x ) 180 × 16 n 1280 f e ( 2 n 1 x ) 180 × 16 n + 4 n 180 f e x 2 n 1 80 × 4 n 180 f e x 2 n + 1024 × 4 n 180 f e x 2 n + 1 + 2 n 5 × 8 n 90 f o x 2 n 1 40 × 2 n 170 × 8 n 90 f o x 2 n + 256 × 2 n 320 × 8 n 90 f o x 2 n + 1 + 4 90 × 32 n f o ( 2 n + 1 x ) 10 f o ( 2 n x ) + 16 f o ( 2 n 1 x ) i f 3 < p < 4 , 4 f e ( 2 n + 1 x ) 180 × 64 n 80 f e ( 2 n x ) 180 × 64 n + 256 f e ( 2 n 1 x ) 180 × 64 n 5 f e ( 2 n + 1 x ) 180 × 16 n + 340 f e ( 2 n x ) 180 × 16 n 1280 f e ( 2 n 1 x ) 180 × 16 n + 4 n 180 f e x 2 n 1 80 × 4 n 180 f e x 2 n + 1024 × 4 n 180 f e x 2 n + 1 + 4 f o ( 2 n + 1 x ) 90 × 32 n 40 f o ( 2 n x ) 90 × 32 n + 64 f o ( 2 n 1 x ) 90 × 32 n 5 f o ( 2 n + 1 x ) 90 × 8 n + 170 f o ( 2 n x ) 90 × 8 n 320 f o ( 2 n 1 x ) 90 × 8 n + 2 n 90 f o x 2 n 1 40 × 2 n 90 f o x 2 n + 256 × 2 n 90 f o x 2 n + 1 i f 2 < p < 3 , f e ( 2 n + 2 x ) 2880 × 64 n 20 f e ( 2 n + 1 x ) 2880 × 64 n + 64 f e ( 2 n x ) 2880 × 64 n 5 f e ( 2 n + 2 x ) 2880 × 16 n + 340 f e ( 2 n + 1 x ) 2880 × 16 n 1280 f e ( 2 n x ) 2880 × 16 n + 4 f e ( 2 n + 2 x ) 320 f e ( 2 n + 1 x ) + 4096 f e ( 2 n x ) 2880 × 4 n + 4 f o ( 2 n + 1 x ) 90 × 32 n 40 f o ( 2 n x ) 90 × 32 n + 64 f o ( 2 n 1 x ) 90 × 32 n 5 f o ( 2 n + 1 x ) 90 × 8 n + 170 f o ( 2 n x ) 90 × 8 n 320 f o ( 2 n 1 x ) 90 × 8 n + 2 n 90 f o x 2 n 1 40 × 2 n 90 f o x 2 n + 256 × 2 n 90 f o x 2 n + 1 i f 1 < p < 2 , f e ( 2 n + 2 x ) 2880 × 64 n 20 f e ( 2 n + 1 x ) 2880 × 64 n + 64 f e ( 2 n x ) 2880 × 64 n 5 f e ( 2 n + 2 x ) 2880 × 16 n + 340 f e ( 2 n + 1 x ) 2880 × 16 n 1280 f e ( 2 n x ) 2880 × 16 n + 4 f e ( 2 n + 2 x ) 320 f e ( 2 n + 1 x ) + 4096 f e ( 2 n x ) 2880 × 4 n + f o ( 2 n + 2 x ) 720 × 32 n 10 f o ( 2 n + 1 x ) 720 × 32 n + 16 f o ( 2 n x ) 720 × 32 n 5 f o ( 2 n + 2 x ) 720 × 8 n + 170 f o ( 2 n + 1 x ) 720 × 8 n 320 f o ( 2 n x ) 720 × 8 n + f o ( 2 n + 2 x ) 40 f o ( 2 n + 1 x ) + 256 f o ( 2 n x ) 180 × 2 n i f p < 1
for all x X and all nonnegative integers n. Then
J n f ( x ) J n + 1 f ( x ) =             4 n 45 4 × 16 n 9 + 64 × 64 n 45 Δ f x 2 n + 3 + 2 n 45 4 × 8 n 9 + 64 × 32 n 45 Γ f x 2 n + 3 i f 6 < p , 4 n 180 16 n 36 Δ f x 2 n + 2 Δ f ( 2 n 1 x ) 45 × 64 n + 1 + 2 n 45 4 × 8 n 9 + 64 × 32 n 45 Γ f x 2 n + 3 i f 5 < p < 6 , 4 n 180 16 n 36 Δ f x 2 n + 2 Δ f ( 2 n 1 x ) 45 × 64 n + 1 + 2 n 90 8 n 18 Γ f x 2 n + 2 2 Γ f ( 2 n 1 x ) 45 × 32 n + 1 i f 4 < p < 5 , 4 n 180 Δ f x 2 n + 2 + Δ f ( 2 n 1 x ) 36 × 16 n + 1 Δ f ( 2 n 1 x ) 45 × 64 n + 1 + 2 n 90 8 n 18 Γ f x 2 n + 2 2 Γ f ( 2 n 1 x ) 45 × 32 n + 1 i f 3 < p < 4 , 4 n 180 Δ f x 2 n + 2 + Δ f ( 2 n 1 x ) 36 × 16 n + 1 Δ f ( 2 n 1 x ) 45 × 64 n + 1 + 2 n 90 Γ f x 2 n + 2 + Γ f ( 2 n 1 x ) 18 × 8 n + 1 2 Γ f ( 2 n 1 x ) 45 × 32 n + 1 i f 2 < p < 3 , Δ f ( 2 n x ) 720 × 4 n + 1 + Δ f ( 2 n x ) 576 × 16 n + 1 Δ f ( 2 n x ) 2880 × 64 n + 1 + 2 n 90 Γ f x 2 n + 2 + Γ f ( 2 n 1 x ) 18 × 8 n + 1 2 Γ f ( 2 n 1 x ) 45 × 32 n + 1 i f 1 < p < 2 , Δ f ( 2 n x ) 720 × 4 n + 1 + Δ f ( 2 n x ) 576 × 16 n + 1 Δ f ( 2 n x ) 2880 × 64 n + 1 Γ f ( 2 n x ) 180 × 2 n + 1 + Γ f ( 2 n x ) 144 × 8 n + 1 Γ f ( 2 n x ) 720 × 32 n + 1 i f p < 1
for all x X and all nonnegative integers n.
Lemma 4.
If f : X Y is a mapping such that D f ( x , y ) = 0 for all x , y X , then J n f ( x ) = f ( x ) for all x X and all positive integers n.
Lemma 5.
If f : X Y is a mapping such that f ( 0 ) = 0 and D f ( x , y ) = 0 for all x , y X \ { 0 } , then D f ( x , y ) = 0 for all x , y X .
Proof. 
Since D f ( x , 0 ) = 0 and D f ( 0 , y ) = D f ( 7 y , y ) = 0 for all x , y X \ { 0 } , the equality D f ( x , y ) = 0 holds for all x , y X . □
From Lemmas 4 and 5, we can prove the following Hyers–Ulam–Rassias stability of the sextic functional equation.
Theorem 3.
Let p 1 , 2 , 3 , 4 , 5 , 6 be a fixed real number. Suppose that f : X Y is a mapping such that:
D f ( x , y ) θ ( x p + y p )
for all x , y X \ { 0 } . Then, there exists a unique general sextic mapping F such that
f ( x ) f ( 0 ) F ( x ) :             K ( 2 p 2 ) + ( 128 + 44 × 2 p ) K ( 2 p 32 ) ( 2 p 8 ) + K ( 2 p 4 ) + ( 256 + 44 × 2 p ) K ( 2 p 64 ) ( 2 p 16 ) θ x p 45 × 2 2 p i f 6 < p , K 2 p ( 2 p 2 ) + ( 128 + 44 × 2 p ) K ( 2 p 32 ) ( 2 p 8 ) 2 p + ( 2 p 1 ) K ( 2 p 4 ) ( 2 p 16 ) + K ( 64 2 p ) θ x p 45 × 2 p i f 5 < p < 6 , ( 2 × 2 p 1 ) K ( 2 p 8 ) ( 2 p 2 ) + 2 K ( 32 2 p ) + ( 2 p 1 ) K ( 2 p 4 ) ( 2 p 16 ) + K ( 64 2 p ) θ x p 45 × 2 p i f 4 < p < 5 , ( 2 × 2 p 1 ) K ( 2 p 8 ) ( 2 p 2 ) + 2 K ( 32 2 p ) + K 4 ( 2 p 4 ) + ( 256 2 p ) K 4 ( 64 2 p ) ( 16 2 p ) θ x p 45 × 2 p i f 3 < p < 4 , K ( 2 p 2 ) + ( 128 2 p ) K ( 32 2 p ) ( 8 2 p ) + K 2 ( 2 p 4 ) + ( 256 2 p ) K 2 ( 64 2 p ) ( 16 2 p ) θ x p 90 × 2 p i f 2 < p < 3 , K 2 p ( 2 p 2 ) + ( 128 2 p ) K ( 32 2 p ) ( 8 2 p ) 2 p + ( 44 + 2 p ) K 32 ( 16 2 p ) ( 4 2 p ) + K 32 ( 64 2 p ) θ x p 90 i f 1 < p < 2 , ( 22 + 2 p ) K ( 8 2 p ) ( 2 2 p ) + K ( 32 2 p ) + ( 44 + 2 p ) K 4 ( 16 2 p ) ( 4 2 p ) + K 4 ( 64 2 p ) θ x p 720 i f p < 1
for all x X \ { 0 } and F ( 0 ) = 0 , where K : = 166 + 43 × 2 p + 112 × 3 p + 6 p and K : = 184 + 57 × 2 p + 112 × 3 p + 6 p .
Proof. 
If f ˜ is the mapping defined by f ˜ ( x ) = f ( x ) f ( 0 ) , then D f ˜ ( x , y ) = D f ( x , y ) and f ˜ ( 0 ) = 0 . By (13) and the definitions of Γ f ˜ and Δ f ˜ , we have:
Γ f ˜ ( x ) = D f o ( 6 x , 2 x ) + 6 D f o ( x , x ) + 42 D f o ( 2 x , x ) + 112 D f o ( 3 x , x ) θ ( 166 + 43 × 2 p + 112 × 3 p + 6 p ) x p , Δ f ˜ ( x ) = D f e ( 6 x , 2 x ) + 8 D f e ( x , x ) + 56 D f e ( 2 x , x ) + 112 D f e ( 3 x , x ) θ ( 184 + 57 × 2 p + 112 × 3 p + 6 p ) x p
for all x X \ { 0 } . It follows from (12) and (13) that
J n f ˜ ( x ) J n + 1 f ˜ ( x ) :             2 n K 45 × 2 ( n + 3 ) p 4 × 8 n K 9 × 2 ( n + 3 ) p + 64 × 32 n K 45 × 2 ( n + 3 ) p + 4 n K 45 × 2 ( n + 3 ) p 4 × 16 n K 9 × 2 ( n + 3 ) p + 64 × 64 n K 45 × 2 ( n + 3 ) p θ x p if 6 < p , 2 n K 45 × 2 ( n + 3 ) p 4 × 8 n K 9 × 2 ( n + 3 ) p + 64 × 32 n K 45 × 2 ( n + 3 ) p + ( 5 × 16 n 4 n ) K 180 × 2 ( n + 2 ) p + 2 × 2 ( n 1 ) p K ) 90 × 64 n + 1 θ x p if 5 < p < 6 , ( 5 × 8 n 2 n ) K 90 × 2 ( n + 2 ) p + 2 × 2 ( n 1 ) p K ) 45 × 32 n + 1 + ( 5 × 16 n 4 n ) K 180 × 2 ( n + 2 ) p + 2 × 2 ( n 1 ) p K ) 90 × 64 n + 1 θ x p if 4 < p < 5 , ( 5 × 8 n 2 n ) K 90 × 2 ( n + 2 ) p + 2 × 2 ( n 1 ) p K ) 45 × 32 n + 1 + 4 n K 180 × 2 ( n + 2 ) p + ( 5 × 4 n + 1 4 ) × 2 ( n 1 ) p K 180 × 64 n + 1 θ x p if 3 < p < 4 , 2 n K 90 × 2 ( n + 2 ) p + ( 5 × 4 n + 1 4 ) × 2 ( n 1 ) p K 90 × 32 n + 1 + 4 n K 180 × 2 ( n + 2 ) p + ( 5 × 4 n + 1 4 ) × 2 ( n 1 ) p K 180 × 64 n + 1 θ x p if 2 < p < 3 , 2 n K 90 × 2 ( n + 2 ) p + ( 5 × 4 n + 1 4 ) × 2 ( n 1 ) p K 90 × 32 n + 1 + 2 n p K 720 × 4 n + 1 2 n p K 576 × 16 n + 1 + 2 n p K 2880 × 64 n + 1 θ x p if 1 < p < 2 , 2 n p K 180 × 2 n + 1 2 n p K 144 × 8 n + 1 + 2 n p K 720 × 32 n + 1 + 2 n p K 720 × 4 n + 1 2 n p K 576 × 16 n + 1 + 2 n p K 2880 × 64 n + 1 θ x p if p < 1
for all x X \ { 0 } . Together with the equality J n f ˜ ( x ) J n + m f ˜ ( x ) = i = n n + m 1 ( J i f ˜ ( x ) J i + 1 f ˜ ( x ) ) for all x X , we obtain that J n f ˜ ( x ) J n + m f ˜ ( x ) :
i = n n + m 1 2 i K 20 × 8 i K + 64 × 32 i K + 4 i K 20 × 16 i K + 64 × 64 i K 45 × 2 ( i + 3 ) p θ x p if 6 < p , i = n n + m 1 2 i K 20 × 8 i K + 64 × 32 i K 45 × 2 ( i + 3 ) p + ( 5 × 16 i 4 i ) K 180 × 2 ( i + 2 ) p + 2 × 2 ( i 1 ) p K ) 90 × 64 i + 1 θ x p if 5 < p < 6 , i = n n + m 1 ( 5 × 8 i 2 i ) K 90 × 2 ( i + 2 ) p + 2 × 2 ( i 1 ) p K ) 45 × 32 i + 1 + ( 5 × 16 i 4 i ) K 180 × 2 ( i + 2 ) p + 2 × 2 ( i 1 ) p K ) 90 × 64 i + 1 θ x p if 4 < p < 5 , i = n n + m 1 ( 5 × 8 i 2 i ) K 90 × 2 ( i + 2 ) p + 2 × 2 ( i 1 ) p K ) 45 × 32 i + 1 + 4 i K 180 × 2 ( i + 2 ) p + ( 5 × 4 i + 1 4 ) × 2 ( i 1 ) p K 180 × 64 i + 1 θ x p if 3 < p < 4 , i = n n + m 1 32 × 64 i K + ( 5 × 4 i + 1 4 ) × 2 ( 2 i + 1 ) p K 90 × 32 i + 1 × 2 ( i + 2 ) p + 4 4 i + 3 + ( 5 × 4 i + 1 4 ) × 2 ( 2 i + 1 ) p K 180 × 64 i + 1 2 ( i + 2 ) p θ x p if 2 < p < 3 , i = n n + m 1 32 × 64 i K + ( 5 × 4 i + 1 4 ) × 2 ( 2 i + 1 ) p K 90 × 32 i + 1 × 2 ( i + 2 ) p + ( 4 i + 2 5 ) 2 i p K 2880 × 16 i + 1 + 2 i p K 2880 × 64 i + 1 θ x p if 1 < p < 2 , i = n n + m 1 2 i p K 180 × 2 i + 1 2 i p K 144 × 8 i + 1 + 2 i p K 720 × 32 i + 1 + ( 4 i + 2 5 ) 2 i p K 2880 × 16 i + 1 + 2 i p K 2880 × 64 i + 1 θ x p if p < 1
for all x X \ { 0 } and n , m N { 0 } . It follows from (15) that the sequence { J n f ˜ ( x ) } is a Cauchy sequence for all x X \ { 0 } . Since Y is complete and f ˜ ( 0 ) = 0 , the sequence { J n f ˜ ( x ) } converges for all x X . Hence, we can define a mapping F : X Y by:
F ( x ) : = lim n J n f ˜ ( x )
for all x X . Moreover, letting n = 0 and passing the limit n in (15), we get the inequality (14). For the case 2 < p < 3 , from the definition of F, we easily get:
D F ( x , y ) = lim n D f e ( 2 n + 1 x , 2 n + 1 y ) 45 × 64 n 20 D f e ( 2 n x , 2 n y ) 45 × 64 n + 64 D f o ( 2 n 1 x , 2 n 1 y ) 45 × 64 n D f e ( 2 n + 1 x , 2 n + 1 y ) 68 D f e ( 2 n x , 2 n y ) + 256 D f e ( 2 n 1 x , 2 n 1 y ) 36 × 16 n + 4 n 180 D f e 2 x 2 n , 2 y 2 n 80 D f e x 2 n , y 2 n + 1024 D f e x 2 n + 1 , y 2 n + 1 + 4 D f o ( 2 n + 1 x , 2 n + 1 y ) 90 × 32 n 40 D f o ( 2 n x , 2 n y ) 90 × 32 n + 64 D f o ( 2 n 1 x , 2 n 1 y ) 90 × 32 n 5 [ D f o ( 2 n + 1 x , 2 n + 1 y ) 34 D f o ( 2 n x , 2 n y ) + 64 D f o ( 2 n 1 x , 2 n 1 y ) ] 90 × 8 n + 2 n 90 D f o 2 x 2 n , 2 y 2 n 40 D f o x 2 n , y 2 n + 256 D f o x 2 n + 1 , y 2 n + 1 lim n ( ( 4 p + 20 × 2 p + 64 ) 2 ( n 1 ) p 45 × 64 n + ( 4 p + 68 × 2 p + 256 ) 2 ( n 1 ) p 36 × 16 n + 4 n ( 4 p + 80 × 2 p + 1024 ) 180 × 2 ( n + 1 ) p + 4 ( 4 p + 10 × 2 p + 16 ) 2 ( n 1 ) p 90 × 32 n + ( 4 p + 34 × 2 p + 64 ) 2 ( n 1 ) p 18 × 8 n + 2 n ( 4 p + 40 × 2 p + 256 ) 90 × 2 ( n + 1 ) p ) × θ ( x p + y p ) = 0
for all x , y X \ { 0 } . Since D F ( x , y ) = 0 for all x , y X \ { 0 } , F : X Y satisfies the equality D F ( x , y ) = 0 for all x , y X by Lemma 5. Furthermore, we easily show that D F ( x , y ) = 0 by a similar method for the other cases, either p < 1 , or 1 < p < 2 , or 3 < p < 4 , or 4 < p < 5 , or 5 < p < 6 , or 6 < p . To prove the uniqueness of F, let F : X Y be another sextic mapping satisfying (14) and F ( 0 ) = 0 . By Lemma 4, the equality F ( x ) = J n F ( x ) holds for all n N . For the case 2 < p < 3 , we have:
J n f ˜ ( x ) F ( x )             = J n f ˜ ( x ) J n F ( x ) ( f ˜ e F e ) ( 2 n + 1 x ) 45 × 64 n + 20 ( f ˜ e F e ) ( 2 n x ) 45 × 64 n + 64 ( f ˜ e F e ) ( 2 n 1 x ) 45 × 64 n + ( f ˜ e F e ) ( 2 n + 1 x ) 36 × 16 n + 68 ( f ˜ e F e ) ( 2 n x ) 36 × 16 n + 256 ( f ˜ e F e ) ( 2 n 1 x ) 36 × 16 n + 4 n 180 ( f ˜ e F e ) 2 x 2 n + 80 ( f ˜ e F e ) x 2 n + 1024 ( f ˜ e F e ) x 2 n + 1 + 4 ( f ˜ o F o ) ( 2 n + 1 x ) 90 × 32 n + 40 ( f ˜ o F o ) ( 2 n x ) 90 × 32 n + 64 ( f ˜ o F o ) ( 2 n 1 x ) 90 × 32 n + 5 ( f ˜ o F o ) ( 2 n + 1 x ) 90 × 8 n + 170 ( f ˜ o F o ) ( 2 n x ) 90 × 8 n + 320 ( f ˜ o F o ) ( 2 n 1 x ) 90 × 8 n + 2 n 90 ( f ˜ o F o ) 2 x 2 n + 40 ( f ˜ o F o ) x 2 n + 256 ( f ˜ o F o ) x 2 n + 1 ( ( 4 p + 20 × 2 p + 64 ) 2 ( n 1 ) p 45 × 64 n + ( 4 p + 68 × 2 p + 256 ) 2 ( n 1 ) p 36 × 16 n + ( 4 p + 80 × 2 p + 1024 ) 2 n 180 × 2 ( n + 1 ) p + 2 ( 4 p + 10 × 2 p + 16 ) 2 ( n 1 ) p 45 × 32 n + ( 4 p + 34 × 2 p + 64 ) 2 ( n 1 ) p 18 × 8 n + ( 4 p + 40 × 2 p + 256 ) 2 n 90 × 2 ( n + 1 ) p ) × K 2 p 2 + ( 128 2 p ) K ( 32 2 p ) ( 8 2 p ) + K 2 ( 2 p 4 ) + ( 256 2 p ) K 2 ( 64 2 p ) ( 16 2 p ) θ x p 90 × 2 p
for all x X \ { 0 } and all positive integers n. Taking the limit in the above inequality as n , we obtain the equality F ( x ) = lim n J n f ˜ ( x ) for all x X \ { 0 } , which means that F ( x ) = F ( x ) for all x X \ { 0 } . Furthermore, we easily show that F ( x ) = F ( x ) by a similar method for the other cases, either p < 1 , or 1 < p < 2 , or 3 < p < 4 , or 4 < p < 5 , or 5 < p < 6 , or 6 < p . □
From Theorem 3, we also prove the hyperstability of the sextic functional equation when p < 0 .
Theorem 4.
Let p < 0 be a real number. If a mapping f : X Y satisfies the inequality ( ) for all x , y X \ { 0 } , then f : X Y is a sextic mapping itself.
Proof. 
According to Theorem 3, there is a unique sextic mapping F of the functional equation D F ( x , y ) = 0 such that:
f ˜ ( x ) F ( x ) ( 22 + 2 p ) K ( 8 2 p ) ( 2 2 p ) + K ( 32 2 p ) + ( 44 + 2 p ) K 4 ( 16 2 p ) ( 4 2 p ) + K 4 ( 64 2 p ) θ x p 720
for all x X \ { 0 } and F ( 0 ) = 0 . From the equality:
D f ˜ ( n x , ( n 1 ) x ) = D f ˜ ( n x , ( n 1 ) x ) D F ( n x , ( n 1 ) x ) = i = 0 7 7 C i ( 1 ) 7 i ( f ˜ F ) ( n x i ( n 1 ) x )
for all x X \ { 0 } and n N , we have the inequality:
7 C 1 ( f ˜ F ) ( x ) = D f ( n x , ( n 1 ) x ) + ( f ˜ F ) ( n x ) i = 2 7 7 C i ( 1 ) 7 i ( f ˜ F ) ( n x i ( n 1 ) x ) θ x p [ n p + ( n 1 ) p + n p + i = 2 7 7 C i ( i ( n 1 ) n ) p × ( 22 + 2 p ) K ( 8 2 p ) ( 2 2 p ) + K ( 32 2 p ) + ( 44 + 2 p ) K 4 ( 16 2 p ) ( 4 2 p ) + K 4 ( 64 2 p ) ]
for all x X \ { 0 } and n N \ { 1 , 2 } . Since f ˜ ( 0 ) = F ( 0 ) and n p , i = 2 7 7 C i ( i ( n 1 ) n ) p , and ( n 1 ) p tends to zero as n , we get f ˜ ( x ) = F ( x ) for all x X . Therefore, D f ( x , y ) = D f ˜ ( x , y ) = D F ( x , y ) = 0 for all x , y X . □

Funding

This research received no external funding.

Conflicts of Interest

The author declares no conflict of interest.

References

  1. Ulam, S.M. A Collection of Mathematical Problems; Interscience: New York, NY, USA, 1960. [Google Scholar]
  2. Hyers, D.H. On the stability of the linear functional equation Proc. Natl. Acad. Sci. USA 1941, 27, 222–224. [Google Scholar]
  3. Rassias, T.M. On the stability of the linear mapping in Banach spaces. Proc. Am. Math. Soc. 1978, 72, 297–300. [Google Scholar]
  4. Gajda, Z. On stability of additive mappings. Int. J. Math. Math. Sci. 1991, 14, 431–434. [Google Scholar]
  5. Găvruta, P. A generalization of the Hyers–Ulam–Rassias stability of approximately additive mappings. J. Math. Anal. Appl. 1994, 184, 431–436. [Google Scholar]
  6. Isac, G.; Rassias, T.M. On the Hyers-Ulam stability of ψ-additive mappings. J. Approx. Theory 1993, 72, 131–137. [Google Scholar]
  7. Jung, S.-M. On the Hyers–Ulam–Rassias stability of approximately additive mappings. J. Math. Anal. Appl. 1996, 204, 221–226. [Google Scholar]
  8. Lee, Y.-H.; Jun, K.W. On the stability of approximately additive mappings. Proc. Am. Math. Soc. 2000, 128, 1361–1369. [Google Scholar]
  9. Gordji, M.E.; Alizadeh, B.; de la Sen, M.; Ghaemi, M.B. Fixed points and approximately C*-ternary quadratic higher derivations. Int. J. Geom. Methods Mod. Phys. 2012, 2012, 1320017. [Google Scholar]
  10. Yang, X.; Shen, G.; Liu, G.; Chang, L. The Hyers–Ulam–Rassias stability of the quartic functional equation in fuzzy β-normed spaces. J. Inequal. Appl. 2015, 2015, 342. [Google Scholar]
  11. Baker, J. A general functional equation and its stability. Proc. Natl. Acad. Sci. USA 2005, 133, 1657–1664. [Google Scholar]
  12. Jun, K.-W.; Kim, H.-M. On the Hyers–Ulam–Rassias stability of a general cubic functional equation. Math. Inequal. Appl. 2003, 6, 289–302. [Google Scholar]
  13. Lee, Y.-H. On the generalized Hyers-Ulam stability of the generalized polynomial function of degree 3. Tamsui Oxf. J. Math. Sci. 2008, 24, 429–444. [Google Scholar]
  14. Lee, Y.-H. On the Hyers–Ulam–Rassias stability of the generalized polynomial function of degree 2. J. Chungcheong Math. Soc. 2009, 22, 201–209. [Google Scholar]
  15. Lee, Y.-H. On the Hyers–Ulam–Rassias stability of a general quartic functional equation. East Asian Math. J. 2019, 35, 351–356. [Google Scholar]
  16. Lee, Y.-H. Stability of a monomial functional equation on a restricted domain. Mathematics 2017, 5, 53. [Google Scholar]

Share and Cite

MDPI and ACS Style

Lee, Y.-H. On the Hyers-Ulam-Rassias Stability of a General Quintic Functional Equation and a General Sextic Functional Equation. Mathematics 2019, 7, 510. https://doi.org/10.3390/math7060510

AMA Style

Lee Y-H. On the Hyers-Ulam-Rassias Stability of a General Quintic Functional Equation and a General Sextic Functional Equation. Mathematics. 2019; 7(6):510. https://doi.org/10.3390/math7060510

Chicago/Turabian Style

Lee, Yang-Hi. 2019. "On the Hyers-Ulam-Rassias Stability of a General Quintic Functional Equation and a General Sextic Functional Equation" Mathematics 7, no. 6: 510. https://doi.org/10.3390/math7060510

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

Article Metrics

Back to TopTop