Stability of additivity and fixed point methods

Introduction In applications quite often we have to do with functions that satisfy some equations only approximately. One of possible ways to deal with them is just to replace such functions by corresponding (in suitable ways) exact solutions to those equations. But there of course arises the issue of errors which we commit in this way. Some tools to evaluate such errors are provided within the theory of Ulam’s type stability. For instance, we can introduce the following definition, which somehow describes the main ideas of such stability notion for the Cauchy equation


Introduction
In applications quite often we have to do with functions that satisfy some equations only approximately. One of possible ways to deal with them is just to replace such functions by corresponding (in suitable ways) exact solutions to those equations. But there of course arises the issue of errors which we commit in this way. Some tools to evaluate such errors are provided within the theory of Ulam's type stability. For instance, we can introduce the following definition, which somehow describes the main ideas of such stability notion for the Cauchy equation f (x + y) = f (x) + f (y). () Definition  Let (A, +) and (X, +) be semigroups, d be a metric in X, E ⊂ C ⊂ R A  + be nonempty, and T be an operator mapping C into R A + (R + stands for the set of nonnegative reals). We say that Cauchy equation () is (E, T )-stable provided for every ε ∈ E and ϕ  ∈ X A with d ϕ  (x + y), ϕ  (x) + ϕ  (y) ≤ ε(x, y), x, y ∈ A, there exists a solution ϕ ∈ X A of equation () such that (As usual, C D denotes the family of all functions mapping a set D = ∅ into a set C = ∅.) Roughly speaking, (E, T )-stability of equation () means that every approximate (in the sense of ()) solution of () is always close (in the sense of ()) to an exact solution to ().
Let us mention that this type of stability has been a very popular subject of investigations for the last nearly fifty years (see, e.g., [-] Hyers [] published a partial answer to it, which can be stated as follows.
Let E and Y be Banach spaces and ε > . Then, for every g : there is a unique f : E → Y that is additive (i.e., satisfies equation ()) and such that

Quite often we describe that result of Hyers simply saying that Cauchy functional equation () is Hyers-Ulam stable (or has the Hyers-Ulam stability).
In the next few years, Hyers and Ulam published some further stability results for polynomial functions, isometries and convex functions in [-]. Let us mention yet that now we are aware of an earlier (than that of Hyers) result concerning such stability that is due to Pólya and Szegö [, Teil I, Aufgabe ] (see also [, Part I, Ch. , Problem ]) and reads as follows (N stands for the set of positive integers).
For every real sequence (a n ) n∈N with sup n,m∈N |a n+ma na m | ≤ , there is a real number ω such that sup n∈N |a n -ωn| ≤ . Moreover, ω = lim n→∞ a n /n. The next theorem is considered to be one of the most classical results.
Theorem  Let E  and E  be two normed spaces, E  be complete, c ≥  and p =  be fixed real numbers. Let f : E  → E  be a mapping such that Then there exists a unique additive function T : The second statement of Theorem , for p < , can be described as the ϕ-hyperstability of the additive Cauchy equation for ϕ(x, y) ≡ c( x p + y p ). Unfortunately, such result does not remain valid if we restrict the domain of f , as the following remark shows it.

The main result
In this paper we prove the following complement to Theorem , which covers also the situation described in Remark  (see Remark ).
Theorem  Let (X, +) be a commutative semigroup, (E, +) be a commutative group, d be a complete metric in E which is invariant (i.e., d(x + z, y + z) = d(x, y) for x, y, z ∈ X), and h : X → R + be a function such that M  := n ∈ N : s(n) + s(n + ) <  = ∅, where s(n) := inf{t ∈ R + : h(nx) ≤ th(x) for all x ∈ X} for n ∈ N. Assume that f : X → E satisfies the inequality Then there exists a unique additive T : It is easily seen that Theorem  yields the subsequent corollary. Assume that f : X → E satisfies (). Then there is a unique additive T :

Corollary  Let X, E and d be as in Theorem
Remark  If E  and E  are normed spaces and X is a subsemigroup of the group (E  , +) such that  / ∈ X, then it is easily seen that the function h : X → E  , given by h(x) = c x p for x ∈ X, with some real p <  and c > , fulfils condition (). This shows that Corollary  (and therefore Theorem , as well) complements Theorem  and in particular also Theorem .
Note that for such h, () takes the form which is sharper than () for p < . A bit more involved example of h : X → E  satisfying () we obtain taking for any real p < , any bounded function γ : X → R with inf γ (X) > , and any A : In some cases, estimation () provided in Corollary  is optimum as the subsequent example shows. Unfortunately, this is not always the case, because the possibly sharpest such estimation we have in Theorem  for p < .
We show that Actually, the calculations are very elementary, but for the convenience of readers, we provide them.
So, fix x, y ∈ X. Suppose, for instance, that A(x) ≤ A(y). Then which means that

Auxiliary result
The [] for a generalization; see also []) and later applied in numerous papers; for a survey on this subject, we refer to [].
We need to introduce the following hypotheses.
(H) X is a nonempty set and (Y , d) is a complete metric space.
(H) f  , f  : X → X are given maps.
(H) T : Y X → Y X is an operator satisfying the inequality (H) : R X + → R X + is an operator defined by Now we are in a position to present the above mentioned fixed point result following from [, Theorem ].
Theorem  Assume that hypotheses (H)-(H) are valid. Suppose that there exist functions ε : X → R + and ϕ : X → Y such that Then there exists a unique fixed point ψ of T with Moreover, ψ(x) := lim n→∞ (T n ϕ)(x) for x ∈ X.

Proof of Theorem 3
Note that () with y = mx gives (   ) http://www.fixedpointtheoryandapplications.com/content/2013/1/285 Define operators T m : E X → E X and m : R X () Then it is easily seen that, for each m ∈ N, := m has the form described in (H) with f  (x) = mx and f  (x) = ( + m)x. Moreover, since d is invariant, () can be written in the form and Now, we can use Theorem  with Y = E and ϕ = f . According to it, the limit exists for each x ∈ X and m ∈ M  , The case j =  is exactly (). So, fix l ∈ N  and assume that () holds for j = l. Then, in view of (), Thus we have shown (). Now, letting j → ∞ in (), we get Thus we have also proved that T m = T m  for each m ∈ M  , which (in view of ()) yields This implies () with T := T m  ; clearly, equality () means the uniqueness of T as well.
Remark  Note that from the above proof we can derive a much stronger statement on the uniqueness of T than the one formulated in Theorem . Namely, it is easy to see that T := T m  is the unique additive mapping such that () holds with some L > .