Loading [MathJax]/jax/output/SVG/jax.js
Research article

Approximation of involution in multi-Banach algebras: Fixed point technique

  • Received: 06 December 2020 Accepted: 24 March 2021 Published: 26 March 2021
  • MSC : 39B52, 39B82, 47H10, 46L05

  • In this research work, we demonstrate the Hyers-Ulam stability for Cauchy-Jensen functional equation in multi-Banach algebras by the fixed point technique. In fact, we prove that for a function which is approximately Cauchy-Jensen in multi Banach algebra, there is a unique involution near it. Next, we show that under some conditions the involution is continuous, the multi-Banach algebra becomes multi-C-algebra and the Banach algebra is self-adjoint.

    Citation: Ehsan Movahednia, Choonkil Park, Dong Yun Shin. Approximation of involution in multi-Banach algebras: Fixed point technique[J]. AIMS Mathematics, 2021, 6(6): 5851-5868. doi: 10.3934/math.2021346

    Related Papers:

    [1] Ehsan Movahednia, Young Cho, Choonkil Park, Siriluk Paokanta . On approximate solution of lattice functional equations in Banach f-algebras. AIMS Mathematics, 2020, 5(6): 5458-5469. doi: 10.3934/math.2020350
    [2] Songkran Pleumpreedaporn, Chanidaporn Pleumpreedaporn, Weerawat Sudsutad, Jutarat Kongson, Chatthai Thaiprayoon, Jehad Alzabut . On a novel impulsive boundary value pantograph problem under Caputo proportional fractional derivative operator with respect to another function. AIMS Mathematics, 2022, 7(5): 7817-7846. doi: 10.3934/math.2022438
    [3] Weerawat Sudsutad, Chatthai Thaiprayoon, Sotiris K. Ntouyas . Existence and stability results for ψ-Hilfer fractional integro-differential equation with mixed nonlocal boundary conditions. AIMS Mathematics, 2021, 6(4): 4119-4141. doi: 10.3934/math.2021244
    [4] Ateq Alsaadi, Manochehr Kazemi, Mohamed M. A. Metwali . On generalization of Petryshyn's fixed point theorem and its application to the product of n-nonlinear integral equations. AIMS Mathematics, 2023, 8(12): 30562-30573. doi: 10.3934/math.20231562
    [5] Ugyen Samdrup Tshering, Ekkarath Thailert, Sotiris K. Ntouyas . Existence and stability results for a coupled system of Hilfer-Hadamard sequential fractional differential equations with multi-point fractional integral boundary conditions. AIMS Mathematics, 2024, 9(9): 25849-25878. doi: 10.3934/math.20241263
    [6] Xiaoming Wang, Rizwan Rizwan, Jung Rey Lee, Akbar Zada, Syed Omar Shah . Existence, uniqueness and Ulam's stabilities for a class of implicit impulsive Langevin equation with Hilfer fractional derivatives. AIMS Mathematics, 2021, 6(5): 4915-4929. doi: 10.3934/math.2021288
    [7] Sabri T. M. Thabet, Sa'ud Al-Sa'di, Imed Kedim, Ava Sh. Rafeeq, Shahram Rezapour . Analysis study on multi-order ϱ-Hilfer fractional pantograph implicit differential equation on unbounded domains. AIMS Mathematics, 2023, 8(8): 18455-18473. doi: 10.3934/math.2023938
    [8] Najla Alghamdi, Bashir Ahmad, Esraa Abed Alharbi, Wafa Shammakh . Investigation of multi-term delay fractional differential equations with integro-multipoint boundary conditions. AIMS Mathematics, 2024, 9(5): 12964-12981. doi: 10.3934/math.2024632
    [9] Choukri Derbazi, Zidane Baitiche, Mohammed S. Abdo, Thabet Abdeljawad . Qualitative analysis of fractional relaxation equation and coupled system with Ψ-Caputo fractional derivative in Banach spaces. AIMS Mathematics, 2021, 6(3): 2486-2509. doi: 10.3934/math.2021151
    [10] J. Vanterler da C. Sousa, E. Capelas de Oliveira, F. G. Rodrigues . Ulam-Hyers stabilities of fractional functional differential equations. AIMS Mathematics, 2020, 5(2): 1346-1358. doi: 10.3934/math.2020092
  • In this research work, we demonstrate the Hyers-Ulam stability for Cauchy-Jensen functional equation in multi-Banach algebras by the fixed point technique. In fact, we prove that for a function which is approximately Cauchy-Jensen in multi Banach algebra, there is a unique involution near it. Next, we show that under some conditions the involution is continuous, the multi-Banach algebra becomes multi-C-algebra and the Banach algebra is self-adjoint.



    Ulam [25] suggested a problem of stability on group homomorphisms in metric groups in 1940. Hyers is the first mathematician who answered the question of Ulam in 1941. He demonstrated the following theorem in [12].

    Theorem 1.1. Let M and N be two Banach spaces and f:MN be a mapping such that

    f(m+n)f(m)f(n)δ

    for some δ>0 and all m,nM. Then there exists a unique additive mapping A:MN such that

    f(m)A(m)δ

    for all mM.

    The results of Hyers' stability theorem were developed by other mathematicians. Recently, numerous consequences concerning the stability of various functional equations in different normed spaces and various control functions have been obtained.

    The problem of stability of some functional equations have been widely explored by direct methods and there are numerous exciting outcomes regarding this problem [1,2,9,13,14,15,16,19,21,23,26]. The fixed point approach has been used to study the Hyers-Ulam stability investigations. The relationship between Hyers-Ulam stability and fixed point theory has been defined in [3,4,20,22,24].

    Definition 1.2. [22] A function d:M×M[0,] is named a generalized metric on the set M if d satisfies the following conditions

    (a) for each m,nM, d(m,n)=0 if and only if m=n;

    (b) for all m,nM, d(m,n)=d(n,m);

    (c) for all m,n,lM, d(m,l)d(m,n)+d(n,l).

    Notice that the just generalized metric significant difference from the metric is that the generalized metric range contains the infinity.

    Theorem 1.3. [22] et (M,d) be a complete generalized metric space and J:MM be a contractive mapping with Lipschitz constant L<1. Then for every mM, either

    d(Js+1m,Jsm)=

    for all nonnegative integers s or there exists an integer s00 so that

    (a) For all ss0, d(Jsm,Js+1m)<;

    (b)Js(m)n, where n is a fixed point of J;

    (c)n is the unique fixed point of J in the set N={nM:d(Js0(m),n)<};

    (d) For each nN, d(n,n)11Ld(Jn,n).

    Dales and Polyakov [8] introduced the concept of multi-normed spaces. We have collected some properties of multi-normed spaces which will be used in this article. We refer readers to [8,17,18] for more details.

    Suppose that (A,.) is a complex normed space and kN. The linear space AA consists of k-tuples (x1,,xk) denote by Ak, where x1,,xkA. The linear operations on Ak are defined coordinate-wise. The zero element of either A or Ak is denoted by 0. The set {1,2,,k} is indicated by Nk and we denote by Σk the group of permutations on k symbols.

    Definition 2.1. [8] A multi-norm on {Ak:kN} is a sequence

    (k)=(k:kN),

    such that k is a norm on Ak for every kN with 1=:

    (M1)(xσ(1),,xσ(k))k=(x1,,xk)k for every σΣk and x1,,xkA.

    (M2)(α1x1,,αkxk)k(maxiNk|αi|)(x1,,xk)k for all α1,,αkC and x1,,xkA.

    (M3) (x1,,xk1,0)k=(x1,,xk1)k1 for all x1,,xk1A.

    (M4) (x1,,xk1,xk1)k=(x1,,xk1)k1 for all x1,,xk1A.

    With these properties, we say that ((Ak,k):kN) is a multi-normed space.

    Lemma 2.2. [18] Assume that ((Ak,k):kN) is a multi-normed space and let kN. Then

    (i)(x,,x)k=x for all xA.

    (ii)maxiNkxi(x1,,xk)kki=1xikmaxiNkxi

    for all x1,,xkA.

    By the second part of the above lemma, we conclude that, if (A,) is a Banach space, then (Ak,k) is a Banach space for every kN. In this case, ((Ak,k):kN) is a multi-Banach space.

    Example 2.3. [8] We define the sequence ((Ak,k):kN) on Ak as follows:

    (x1,,xk)k:=maxiNkxifor all(x1,,xkA).

    This is a minimum multi-norm on Ak.

    Definition 2.4. [8,17] Suppose that (A,) is a normed algebra such that ((Ak,k):kN) is a multi-normed space. Then ((Ak,k):kN) is called a multi-normed algebra if

    (a1b1,,akbk)k(a1,,ak)k(b1,,bk)k,

    for all kN and a1,,ak,b1,,bkA. Moreover, the multi-normed algebra ((Ak,k):kN) is a multi-Banach algebra if ((Ak,k):kN) is a multi-Banach space.

    It is obvious that every Banach algebra is a multi-Banach algebra with minimum multi-norm.

    A mapping :AA, denoted by

    (x)=x,

    is an involution on A if

    (1) x=x,

    (2) (λx+μy)=¯λx+¯μy, for all x,yA and λ,μC,

    (3)(xy)=yx, for all x,yA.

    Definition 2.5. Assume that ((Ak,k):kN) is a multi-Banach algebra. A multi-C-algebra is a complex multi-Banach algebra ((Ak,k):kN) with an involution satisfying

    (a1a1,,akak)k=(a1,,ak)2k,

    for all kN and a1,,akA.

    In all sections of this article, wherever needed, we assume that

    S1:={zC:|z|=1},S11n0:={eiθ;0θ2πn0}

    for n0N. It is obvious that S1=S111. Furthermore, suppose ((Ak,.k):kN) is a multi-Banach algebra. Let I:=2+αα for real α greater than or equal to 2. When A is a Banach algebra, for a mapping f:AA, we define

    Eμηνf(x,y,z)=α¯μf(x+yα+z)f(μx)f(ηy)αf(νz),

    for all x,y,zA and μ,ν,ηS11n0.

    An algebra A is called C-algebra if it is a Banach algebra with an involution such that xx=x2. We suggest [6] as a reference for stability in Banach algebras and C-algebras.

    Lemma 3.1. Let I be a positive real number and φ:A3k[0,) be a function such that there exists L<1 satisfying

    I1φ(Ix1,Iy1,Iz1,,Ixk,Iyk,Izk)Lφ(x1,y1,z1,,xk,yk,zk) (3.1)

    for all x1,x2,,xk,y1,y2,,yk,z1,z2,,zkA. Then

    limjIjφ(Ijx1,Ijy1,Ijz1,,Ijxk,Ijyk,Ijzk)=0.

    Proof. Putting xi=Ixi, yi=Iyi and zi=Izi for 1ik in (3.1), we have

    I1φ(Ix1,Iy1,Iz1,,Ixk,Iyk,Izk)Lφ(x1,y1,z1,,xk,yk,zk).

    Replacing xi,yi,zi by Ixi,Iyi,Izi, respectively, in the above inequality, we get

    I2φ(I2x1,I2y1,I2z1,,I2xk,I2yk,I2zk)L2φ(x1,y1,z1,,xk,yk,zk).

    By induction, we obtain

    Inφ(Inx1,Iny1,Inz1,,Inxk,Inyk,Inzk)Lnφ(x1,y1,z1,,xk,yk,zk).

    Since L<1, we get

    limnInφ(Inx1,Iny1,Inz1,,Inxk,Inyk,Inzk)=0,

    as desired.

    Theorem 3.2. Let ((An,.n):nN) be a multi-Banach algebra and f:AA be a mapping and φ:A3k[0,) be a function satisfying (3.1) and

    (Eμηνf(x1,y1,z1),,Eμηνf(xk,yk,zk))kφ(x1,y1,z1,,xk,yk,zk), (3.2)
    (f(x1y1)f(y1)f(x1),,f(xkyk)f(yk)f(xk))kφ(x1,y1,0,,xk,yk,0), (3.3)
    limmImf(ImlimnInf(Inx))=x (3.4)

    for all x,x1,y1,z1,,xk,yk,zkA and μ,η,νS11n0. Then there exists a unique involution mapping F:AA which satisfies

    (F(x1)f(x1),,F(xk)f(xk))k1(1L)(2+α)φ(x1,x1,x1,,xk,xk,xk). (3.5)

    Furthermore, (a) if

    |(f(x1),f(x2),,f(xk))k(x1,x2,,xk)k|φ(x1,x1,x1,,xk,xk,xk) (3.6)

    for all x1,x2,,xnA, then the involution mapping F:AA is continuous;

    (b) if

    |(x1f(x1),,xkf(xk))k(x1,,xk)2k|φ(x1,x1,x1,,xk,xk,xk) (3.7)

    for all x1,x2,,xkA, then A is a C-algebra with involution x=F(x) for all xA.

    Proof. Let (Δ,d) be a generalized metric space given by the following definition:

    Δ={g:AA;g(0)=0},d(g,h)=inf{c[0,):(g(x1)h(x1),,g(xk)h(xk))kcφ(x1,x1,x1,,xk,xk,xk)} (3.8)

    for all x1,x2,xkA. We can show that the metric space (Δ,d) is complete (see [5, Theorem 2.5]). Next, we define an operator

    Υ:ΔΔ(Υg)(x):=I1g(Ix), (3.9)

    for all xA. We claim that the operator Υ is strictly contractive on Δ. For this, assume that d(g,h)=κ for any g,hΔ, for some κR, we have

    (g(x1)h(x1),,g(xk)h(xk))kκφ(x1,x1,x1,,xk,xk,xk),

    for all x1,x2,,xkA. Replacing x1,x2,,xk by Ix1,Ix2,,Ixk, respectively, in the above inequality, we get

    (g(Ix1)h(Ix1),,g(Ixk)h(Ixk))kκφ(Ix1,Ix1,Ix1,,Ixk,Ixk,Ixk)

    and by (3.1), we obtain

    (I1g(Ix1)I1h(Ix1),,I1g(Ixk)I1h(Ixk))kκI1φ(Ix1,Ix1,Ix1,,Ixk,Ixk,Ixk)=κLφ(x1,x1,x1,,xk,xk,xk).

    So d(Υg,Υh)κL. Hence we get

    d(Υg,Υh)Ld(g,h).

    Then Υ is strictly contactive on Δ by Lipschitz constant L<1.

    Letting μ=η=ν=1 and yi=zi:=xi for 1ik in (3.2), we obtain

    (Eμηνf(x1,x1,x1),,Eμηνf(xk,xk,xk))kφ(x1,x1,x1,,xk,xk,xk), (3.10)

    for all x1,x2,,xkA. Note that

    Eμ,η,νf(x1,x1,x1)=αf(x1+x1α+x1)f(x1)f(x1)αf(x1)=f(2+ααx1)(2+α)f(x1).

    By (3.10), we have

    (αf(2+ααx1)(2+α)f(x1),,αf(2+ααxk)(2+α)f(xk))kφ(x1,x1,x1,,xk,xk,xk)

    for all x1,x2,,xkA. Thus we get

    (I1f(Ix1)f(x1),,I1f(Ixk)f(xk))k1α+2φ(x1,x1,x1,,xk,xk,xk)

    for all x1,x2,,xkA. It follows from (3.9) that

    (Υf(x1)f(x1),,Υf(xk)f(xk))k1α+2φ(x1,x1,x1,,xk,xk,xk),

    for all x1,x2,,xkA, and by (3.8), we get

    d(Υf,f)1α+2<. (3.11)

    The conditions of the fixed point theorem are estabilished, so there exists a mapping F:AA such that F(0)=0 and the following conditions hold.

    (ⅰ) F is a fixed point of Υ. This means that

    (ΥF)(Ix)=F(x)I1F(Ix)=F(x)F(Ix)=IF(x)

    and thus

    F(α+2αx)=α+2αF(x)

    for all xA. Moreover, the mapping F is a unique fixed point in the set

    χ={gΔ:d(g,f)<}.

    Note that n0=0 in Theorem 1.3. From (3.8), there is c[0,) satisfying

    (F(x1)f(x1),,F(xk)f(xk))kcφ(x1,x1,x1,,xk,xk,xk).

    (ⅱ) The sequence {Υnf} converges to F. This implies that

    F(x)=limnInf(Inx)orF(x)=limn(α2+α)nf((2+αα)nx). (3.12)

    (ⅲ) We obtain

    d(F,f)11Ld(ΥF,f)1(1L)(2+α)

    and so (3.5) holds. By Lemma 3.1, (3.1), (3.2) and (3.11), we have

    (αF(x1+y1α+z1)F(x1)F(y1)αF(z1),,αF(xk+ykα+zk)F(xk)F(yk)αF(zk))k=(αlimnInf(In(x1+y1α+z1))limnInf(Inx1)limnInf(Iny1)αlimnInf(Inz1),,αlimnInf(In(xk+ykα+zk))limnInf(Inxk)limnInf(Inyk)αlimnInf(Inzk))k=limnIn[(αf(Inx1+Iny1α+Inz1)f(Inx1)f(Iny1)αf(Inz1),,αf(Inxk+Inykα+Inzk)f(Inxk)f(Inyk)αf(Inzk))k]limnInφ(Inx1,Iny1,Inz1,,Inxk,Inyk,Inzk)=0(by Lemma 3.1).

    So we obtain

    (αF(x1+y1α+z1)F(x1)F(y1)αF(z1),,αF(xk+ykα+zk)F(xk)F(yk)αF(zk))k=0.

    Replacing x1,x2,,xk and y1,y2,,yk and z1,z2,,zk with x,y,z, respectively, we have

    αF(x+yα+z)F(x)F(y)αF(z)=0

    and thus

    αF(x+yα+z)=F(x)+F(y)+αF(z)

    and so F is an additive mapping, that is,

    F(x+y)=F(x)+F(y) (3.13)

    for all x,yA (see [11]).

    We will show that F(xy)=F(y)F(x) for all x,yA. Substituting x1,x2,.,xk with Inx1,Inx2,,Inxk and y1,y2,,yk with Iny1,Iny2,,Inyk in (3.3) and dividing on I2n, we obtain

    2n(f(I2nx1y1)f(Iny1)f(Inx1),,f(I2nxkyk)f(Inyk)f(Inxk))kI2nφ(Inx1,Iny1,0,,Inxk,Inyk,0).

    Letting n, we have

    F(x1y1)F(y1)F(x1),,F(xkyk)F(yk)F(xk)limnI2nφ(Inx1,Iny1,0,,Inxk,Inyk,0)=0(by Lemma 3.1).

    Replacing x1,x2,,xk and y1,y2,,yk with x,y, respectively in the above inequality, we get

    F(xy)F(y)F(x)=0F(xy)=F(y)F(x) (3.14)

    for all x,yA. We will show that F(μx)=¯μF(x) for all xA and all μS11n0.

    The method applied in [10] is used to continue the proof. Substituting xi=yi=zi:=Inxi for every 1ik and η=ν:=μ, we obtain

    (Eμηνf(Inx1,Inx1,Inx1),,Eμηνf(Inxk,Inxk,Inxk))k=(α¯μf(Inx1+Inx1α+Inx1)f(μInx1)f(μInx1)αf(μInx1),,α¯μf(Inxk+Inxkα+Inxk)f(μInxk)f(μInxk)αf(μInxk))k=(α¯μf(In+1x1)(α+2)f(μInx1),,¯μf(In+1xk)(α+2)f(μInxk))kφ(Inx1,Inx1,Inx1,,Inxk,Inxk,Inxk).

    In particular, if μ=1, then

    (αf(In+1x1)(α+2)f(Inx1),,f(In+1xk)(α+2)f(Inxk))kφ(Inx1,Inx1,Inx1,,Inxk,Inxk,Inxk).

    We have

    ((2+α)f(μInx1)(2+α)¯μf(Inx1),,(2+α)f(μInxk)(2+α)¯μf(Inxk))k=((2+α)f(μInx1)α¯μf(In+1x1)+α¯μf(In+1x1)(2+α)¯μf(Inx1),,(2+α)f(μInxk)α¯μf(In+1xk)+α¯μf(In+1xk)(2+α)¯μf(Inxk))k((2+α)f(μInx1)α¯μf(In+1x1),,(2+α)f(μInxk)α¯μf(In+1xk))k+(α¯μf(In+1x1)(2+α)¯μf(Inx1),,α¯μf(In+1xk)(2+α)¯μf(Inxk))k((2+α)f(μInx1)α¯μf(In+1x1),,(2+α)f(μInxk)α¯μf(In+1xk))k+|μ|(αf(In+1x1)(2+α)f(Inx1),,αf(In+1xk)(2+α)f(Inxk))k2φ(Inx1,Inx1,Inx1,,Inxk,Inxk,Inxk)

    for all x1,x2,,xkA. This implies that

    (Inf(μInx1)¯μInf(Inx1),,Inf(μInxk)¯μInf(Inxk))kIn(f(μInx1)¯μf(Inx1),,f(μInxk)¯μf(Inxk))k22+αInφ(Inx1,Inx1,Inx1,,Inxk,Inxk,Inxk)Inφ(Inx1,Inx1,Inx1,,Inxk,Inxk,Inxk)

    for all x1,x2,,xkA. As n, we have

    limn(Inf(Inx1)In¯μf(Inx1),,Inf(Inxk)In¯μf(Inxk))k=0,

    by Lemma 3.1. Therefore,

    (F(μx1)¯μF(x1),,F(μxk)¯μF(xk))k=0.

    Letting x1=x2==xk=x, we get

    F(μx)=¯μF(x) (3.15)

    for all xA and μS11n0. Now, we will show that the above equality holds for all μC. To prove this, we can consider the following statements:

    (1) If μS1 then there exists θ[0,2π] such that μ=eiθ. Putting μ1=eiθn0S11n0, by (3.15), we get

    μn01=μF(μx)=¯μ1n0F(x)=¯μF(x)

    for all xA.

    (2) If μnS1={nz;zS1} for some nN, then by additivity of F, we have

    F(μx)=F(nzx)=¯zF(nx)=n¯zF(x)=¯μF(x).

    (3) Let t(0,). Then there exists a positive integer n such that the point (t,0) lies in the interior of circle with center at origin and radius n. Putting t1:=t+in2t2 and t2:=tin2t2, we have t=t1+t22 and t1,t2nS1. It follows that

    F(tx)=F(t1+t22x)=F(t12x+t22x)=F(t12x)+F(t22x)=¯t12F(x)+¯t22F(x)=t1+t22F(x)=tF(x)

    for all xA. On the other hand, for any λC, there is θ[0,2π] such that λ=|λ|eiθ. It follows that

    F(λx)=F(|λ|eiθx)=|λ|e¯iθF(x)=|λ|eiθF(x)=¯λF(x)

    for all xA. Hence, for any case, we get

    F(λx)=¯λF(x) (3.16)

    for all λC. Using the assumption (3.4), we have

    F(F(x))=F(limnInf(Inx))=limmImf(ImlimnInf(Inx))=x

    and thus

    F(F(x))=x. (3.17)

    Hence F(x) is an involution for A, since

    F(x+y)=F(x)+F(y),i.e.,(x+y)=x+yby (3.13),F(xy)=F(y)F(x),i.e.,(xy)=yxby (3.14),F(λx)=¯λF(x),i.e.,(λx)=¯λxby (3.16),F(F(x))=x,i.e.,(x)=xby (3.17).

    Therefore A is a Banach algebra with an involution F(x). To prove (a) of Theorem 3.2, replacing x1,x2,,xk by Inx1,Inx2,,Inxk in (3.6) and dividing both sides on In, we have

    In|(f(Inx1),f(Inx2),,f(Inxk))k(Inx1,Inx2,,Inxk)k|Inφ(Inx1,Inx1,Inx1,,Inxk,Inxk,Inxk).

    Letting n, we have

    |(F(x1),F(x2),,F(xk))k(x1,x2,,xk)k|limnInφ(Inx1,Inx1,Inx1,,Inxk,Inxk,Inxk)=0.

    Replacing x1,x2,,xk by x, we get

    |F(x)x|=0F(x)=x.

    This implies that the involution F(x) is isometric. Let {xn} be an arbitrary sequence in A such that xnx. Then

    xnx=F(xnx)=F(xn)F(x),

    which indicates that the involution F(x) is continuous. To show that A is a C-algebra, we must prove that

    xF(x)=x2.

    For this purpose, we use the assumption (3.7) and the equality (3.12). If

    |(x1f(x1),,xkf(xk))k(x1,x2,,xk)2k|φ(x1,x1,x1,,xk,xk,xk),

    for all x1,x2,,xkA, then by replacing x1,x2,,xk by Inx1,Inx1,,Inxk and dividing both sides of the above inequality by I2n, we obtain

    I2n|(Inx1f(Inx1),,Inxkf(Inxk))k(Inx1,Inx2,,Inxk)2k|I2nφ(Inx1,Inx1,Inx1,,Inxk,Inxk,Inxk).

    Thus

    |(Inx1f(Inx1),,Inxkf(Inxk))k(x1,x2,,xk)2k|I2nφ(Inx1,Inx1,Inx1,,Inxk,Inxk,Inxk).

    Letting n, we get

    |(x1F(x1),,xkF(xk))k(x1,x2,,xk)2k|limnI2nφ(Inx1,Inx1,Inx1,,Inxk,Inxk,Inxk)=0.

    Therefore,

    (x1F(x1),,xkF(xk))k=(x1,x2,,xk)2k.

    Replacing x1,x2,,xk by x in last inequality, we have

    xF(x)=x2

    for all xA. Then A is a C-algebra with F(x) as an involution.

    Corollary 3.3. Let f:AA be a mapping such that

    (Eμηνf(x1,y1,z1),,Eμηνf(xk,yk,zk))kϵ(ki=1xip+yip+zip),
    (f(x1y1)f(y1)f(x1),,f(xkyk)f(yk)f(xk))kϵ(ki=1xip+yip),
    limmImf(ImlimnInf(Inx))=x

    for all x,x1,x2,,xk,y1,y2,,yk,z1,z2,,zkA and μ,η,νS11n0, p[0,1) and ϵ[0,). Then there is a unique involution mapping F:AA such that

    (F(x1)f(x1),,F(xk)f(xk))k3ϵα(IIp)ki=1xip.

    Furthermore, (a) if

    (f(x1)x1,,f(xk)xk)kϵki=1xip, (3.18)

    for all x1,x2,,xnA, then A=Asa, which means that for every xA, we have x=F(x);(b) if

    |(x1f(x1),,xkf(xk))k(x1,,xk)2k|3ϵki=1xip,

    for all x1,x2,,xkA, then A is a C-algebra with involution x=F(x) for all xA.

    Proof. By taking

    φ(x1,y1,z1,,xk,yk,zk)=ki=1(xip+yip+zip),

    in Theorem 3.2, this corollary is proved. Note that

    limjIjφ(Ijx1,Ijy1,Ijz1,,Ijxk,Ijyk,Ijzk)=limjI(p1)jki=1(xip+yip+zip)=0.

    Furthermore, putting Inxi by xi for all 1ik in (3.18), we get

    (f(Inx1)Inx1,f(Inx2)Inx2,,f(Inxk)Inxk)kϵki=1Inxip.

    Dividing both sides of the above inequality by In, we obtain

    (Inf(Inx1)x1,Inf(Inx2)x2,,Inf(Inxk)xk)kϵIn(1p)ki=1xip.

    Letting n, we have

    (F(x1)x1,F(x2)x2,,F(xk)xk)k=0.

    Replacing x1,x2,,xk by x in the above equality, we obtain

    F(x)=x(i.e.,x=x)

    for all xA. The rest of the proof is similar to the previous one.

    Theorem 3.4. Suppose that f:AA is a mapping which satisfies (3.2), (3.3) and (3.4). Let ψ:A3k[0,) be a function such that there is L<1 satisfying

    ψ(x1,y1,z1,,xk,yk,zk)I2Lψ(Ix1,Iy1,Iz1,,Ixk,Iyk,Izk),

    for all x1,x2,,xk,y1,y2,,yk,z1,z2,,zkA and μ,η,νS11n0. Then there exists a unique involution mapping F:AA which satisfies

    (F(x1)f(x1),,F(xk)f(xk))kLI2α(1L)ψ(x1,x1,x1,,xk,xk,xk).

    Furthermore, (a) if

    (f(x1)x1,,f(xk)xk)kψ(x1,x1,x1,,xk,xk,xk)

    for all x1,x2,,xnA, then A=Asa, that is, for every xA, we have x=F(x); (b) if

    |(x1f(x1),,xkf(xk))k(x1,,xk)2k|ψ(x1,x1,x1,,xk,xk,xk)

    for all x1,x2,,xnA, then A is C-algebra with x=F(x) as an involution.

    Proof. The proof of this theorem is the same as in the previous one. We prove some of the parts and refer the rest on to the readers.

    The linear mapping Υ:ΔΔ is defined by

    Υg(x):=Ig(I1x),

    for all xA. It is easy to show that the operator Υ is strictly contractive. Putting μ=η=ν=1 and replacing xi,yi,zi by Iixi in (3.2) for all 1ik, we get

    (Eμηνf(I1x1,I1x1,I1x1),,Eμηνf(I1xk,I1xk,I1xk))k=(αf(x1)(2+α)f(I1x1),,αf(xk)(2+α)f(I1xk))kψ(I1x1,I1x1,I1x1,,I1xk,I1xk,I1xk).

    Therefore,

    (f(x1)If(I1x1),,f(xk)If(I1xk))k1αψ(I1x1,I1x1,I1x1,,I1xk,I1xk,I1xk),1αI2ψ(x1,x1,x1,,xk,xk,xk)

    for all x1,x2,,xkA. Hence

    (f(x1)Υf(x1),,f(xk)Υf(xk))k1αI2ψ(x1,x1,x1,,xk,xk,xk)

    for all x1,x2,,xkA. We obtain

    d(f,Υf)LI2α.

    There exists a mapping F:AA with F(0)=0, which is a fixed point of Υ and ΥnfF, that is,

    F(x)=limnInf(Inx).

    The remainder is similar to the proof of Theorem 3.2.

    Theorem 3.5. Let A be a Banach algebra and ((Akk):kN) be a multi-Banach algebra. Let pN with p2 and assume that there are 0L1 and functions Γ,φ:A2k[0,) satisfying

    φ(x1,y1,,xk,yk)Lpφ(px1,py1,,pxk,pyk),
    limn1p2nΓ(x1pn,y1pn,,xkpn,ykpn)=0

    for all x1,,xk,y1,,ykA. Assume that f:AA is a mapping satisfying f(0)=0 and

    (f(μx1+νy1)ˉμf(x1)ˉνf(y1),,f(μxk+νyk)ˉμf(xk)ˉνf(yk))kφ(x1,y1,,xk,yk), (3.19)
    (f(x1y1)f(y1)f(x1),,f(xkyk)f(yk)f(xk))kΓ(x1,y1,,xk,yk), (3.20)
    limnpnf(1pnlimmpmf(1pmx))=x (3.21)

    for all μ=1,i and ν=1,i and for all x,x1,x2,,xk, y1,y2,,ykA. Suppose that for any fixed xA the function tf(tx) is continuous on R. Then there is a unique involution T:AA such that

    (f(x1)T(x1),,f(xk)T(xk))k11LΨ(x1,x2,,xk),

    where

    Ψ(x1,x2,,xk):=p1j=1φ(jx1p,x1p,,jxkp,xkp)

    for all x1,x2,,xkA.

    Furthermore, if

    |(x1f(x1),,xkf(xk))k(x1,,xk)2k|Γ(x1,x1,,xk,xk)

    for all x1,x2,,xkA, then A is a C-algebra with involution x=T(x) for all xA. Also, if

    |(f(x1),,f(xk))k(x1,,xk)k|φ(x1,x1,,xk,xk)

    for all x1,x2,,xkA, then the involution mapping T:AA is continuous.

    Proof. Putting μ=ν=1 and y1=x1,y2=x2,,yk=xk in (3.19), we have

    (f(2x1)2f(x1),,f(2xk)2f(xk))kφ(x1,x1,,xk,xk),

    for all x1,x2,,xkA. By using induction, we get

    (f(px1)pf(x1),,f(pxk)pf(xk))kp1j=1φ(jx1,x1,,jxk,xk).

    Replacing xj by xjp in the above inequality for 1jk, we get

    (f(x1)pf(x1p),,f(xk)pf(xkp)r)kΨ(x1,,xk). (3.22)

    Set Δ:={g:AA,g(0)=0} and define

    d:Δ×Δ[0,),d(g,h)=inf{c>0,(g(x1)h(x1),,g(xk)h(xk))kcΨ(x1,,xk)}

    for all x1,x2,,xkA. We can easily show that (Δ,d) is a complete metric space. We assert that the mapping J:ΔΔ defined by Jg(x)=pg(xp) is a strictly contractive mapping. Let g,hΔ with d(g,h)< and d(g,h)<c. Thus we have

    (g(x1)h(x1),,g(xk)h(xk))kcΨ(x1,,xk)

    for all x1,x2,,xkA. Therefore,

    (pg(x1p)ph(x1p),,pg(xkp)ph(xkp))kpcΨ(1px1,,1pxk)cLψ(x1,,xk).

    Thus d(Jg,Jh)cL and so d(Jg,Jh)Ld(g,h) for all g,hΔ. Hence J is a strictly contractive mapping with Lipschitz constant L. By (3.22), we have

    d(f,Jf)1<. (3.23)

    The conditions of the fixed point theorem are satisfied. There exists n0N such that the sequence {Jnf} converges to a fixed point T of J and therefore T(xp)=1pT(x). Also T is a unique fixed point of J in the set χ={gΔ:d(Jn0f,g)<} and

    d(g,T)d(g,Jf)1L(gχ).

    Since limnd(Jnf,T)=0, we have

    limnpnf(xpn)=T(x)(xA).

    From (3.23), we have d(f,T)11L. Hence

    (f(x1)T(x1),,f(xk)T(xk))k11Lφ(x1,x1,,xk,xk).

    Replacing x1,x2,,xk by xpn and y1,y2,,yk by ypn and multiplying both sides by pn, we get

    (pnf(μx1pn+νy1pn)pnˉμf(x1pn)pnˉνf(y1pn),,pnf(μxkpn+νykpn)pnˉμf(xkpn)pnˉνf(ykpn))kpnφ(x1pn,y1pn,,xkpn,ykpn).

    Letting n, we obtain

    (T(μx1+νy1)ˉμT(x1)ˉνT(y1),,T(μxk+νyk)ˉμT(xk)ˉνT(yk))klimnpnφ(x1pn,y1pn,,xkpn,ykpn)=0.

    Therefore,

    (T(μx1+νy1)ˉμT(x1)ˉνT(y1),,T(μxk+νyk)ˉμT(xk)ˉνT(yk))k=0.

    Replacing x1,x2,,xk by x and y1,y2,,yk by y, we have

    T(μx+νy)=ˉμT(x)+ˉνT(y) (3.24)

    for all x,yA and μ,ν=1,i.

    Next, putting x1:=x1pn,,xk:=xkpn and y1:=y1pn,,yk:=ykpn in (3.20), we obtain

    (f(x1pny1pn)f(y1pn)f(x1pn),,f(xkpnykpn)f(ykpn)f(xkpn))kΓ(x1pn,y1pn,,xkpn,ykpn).

    Multiplying both sides in p2n, we get

    (p2nf(x1y1p2n)pnf(y1pn)pnf(x1pn),p2nf(xkykp2n)pnf(ykpn)pnf(xkp))kp2nΓ(x1pn,y1pn,,xkpn,ykpn).

    Letting n in the above inequality, we have

    T(x1y1)T(x1)T(y1),,T(xkyk)T(xk)T(yk)klimnp2nΓ(x1pn,y1pn,,xkpn,ykpn)=0.

    Therefore, we get

    T(xy)=T(x)T(y) (3.25)

    for all x,yA. We must show that T is linear. For this, we can use the method of [7]. Fix x0A and ρA(dual space of A). We define the mapping Φ;RR as follows:

    Φ(t):=ρ(T(tx0))=limnpnρ(f(pntx0)).

    So

    (i)Φ is an additive mapping, i.e., for any a,bR, we have

    Φ(a+b)=Φ(a)+Φ(b).

    It is easy to show that Φ is additive by (3.24).

    (ii)Φ is a Borel function. Put Φn(t)=pnρ(f(pntx0)). Then Φn(t) are continuous, and Φ(t) is the pointwise limit of Φn(t) and thus Φ(t) is a Borel function.

    Also, we know that if Φ:RnRn is a function such that Φ is additive and measurable, then Φ is continuous. Note that, if we replace Rn by any separable, locally compact abelian group and this statement is true. Therefore Φ(t) is a continuous function. Let aR. Then a=limnrn, where {rn} is a sequence of rational numbers. Hence

    Φ(at)=Φ(tlimnrn)=limn(Φ(trn))=limnrnΦ(t)=aΦ(t).

    Thus Φ is R-linear and therefore T(ax)=aT(x) for all aR. For ϱ=a1+ia2C, where a1,a2R, we have

    T(ϱx)=T(a1x+ia2x)=T(a1x)+T(ia2x)=T(a1x)+¯iT(a2x)=T(a1x)iT(a2x)=a1T(x)ia2T(x)=¯ϱT(x).

    Then T is a C-linear mapping. Using the assumption (3.21), we have

    T(T(x))=T(limmpmf(pmx))=limnpnf(pnlimmpmf(pmx))=x. (3.26)

    Now, we can say that T(x) is an involution for A (see (3.24), (3.25), (???) and (3.26)).

    In this research work, we demonstrated the stability for Cauchy-Jensen functional equation in multi-Banach algebra by using the fixed point technique. In fact, we proved that for a function which is approximately Cauchy-Jensen in multi-Banach algebra, there is a unique involution near it. Next, we showed that under some conditions the involution is continuous, the multi-Banach algebra becomes multi-C-algebra and the Banach algebra is self-adjoint.

    We would like to express our sincere gratitude to the anonymous referee for his/her helpful comments that will help to improve the quality of the manuscript.

    The authors equally conceived of the study, participated in its design and coordination, drafted the manuscript, participated in the sequence alignment, and read and approved the final manuscript.

    All authors declare no conflicts of interest in this paper.



    [1] N. K. Agbeko, Stability of maximum preserving functional equation on Banach lattice, Miskolc Math. Notes, 13 (2012), 187–196. doi: 10.18514/MMN.2012.570
    [2] Y. Aissi and D. Zeglami, D'Aembert's matrix functional equation with an endomorphism on Abelian groups, Results Math., 75 (2020), 1–17. doi: 10.1007/s00025-019-1126-4
    [3] J. Brzdkek, L. Cadariu, K. Cieplinski, Fixed point theory and the Ulam stability, J. Funct. Spaces, 2014 (2014), 829419.
    [4] L. Cadariu, V. Radu, Fixed points and the stability of Jensen's functional equation, J. Inequal. Pure Appl. Math., 4 (2003), 4.
    [5] L. Cǎdariu, V. Radu, On the stability of the Cauchy functional equation: a fixed point approach, Grazer Math. Ber., 346 (2004), 43–52.
    [6] Y. Cho, C. Park, Th. M. Rassias, R. Saadati, Stability of Functional Equations in Banach Algebras, Springer, New York, 2015.
    [7] H. G. Dales, M. S. Moslehian, Stability of mappings on multi-normed spaces, Glasgow Math. J., 49 (2007), 321–332. doi: 10.1017/S0017089507003552
    [8] H. G. Dales, M. E. Polyakov, Multi-normed spaces and multi-Banach algebras, preprint.
    [9] I. El-Fassi, Solution and approximation of radical quintic functional equation related to quintic mapping in quasi-β-Banach spaces, Rev. R. Acad. Cienc. Exactas Fís. Nat. Ser. A Mat. RACSAM, 113 (2019), 675–687. doi: 10.1007/s13398-018-0506-z
    [10] M. Eshaghi Gordji, Nearly involutions on Banach algebras: A fixed point approach, Fixed Point Theory, 14 (2013), 117–124.
    [11] Z. X. Gao, H. X. Cao, W. T. Zheng, L. Xu, Generalized Hyers-Ulam-Rassias stability of functional inequalities and functional equations, J. Math. Inequal., 3 (2009), 63–77.
    [12] H. D. Hyers, On the stability of the linear functional equation, Proc. Nat. Acad. Sci. U.S.A., 27 (1941), 222–224. doi: 10.1073/pnas.27.4.222
    [13] S. Jung, Hyers-Ulam-Rassias stability of Jensen's equation and its application, Proc. Am. Math. Soc., 126 (1998), 3137–3143. doi: 10.1090/S0002-9939-98-04680-2
    [14] S. Jung, D. Popa, M. Th. Rassias, On the stability of the linear functional equation in a single variable on complete metric spaces, J. Global Optim., 59 (2014), 13–16.
    [15] Y. Lee, S. Jung, M. Th. Rassias, Uniqueness theorems on functional inequalities concerning cubic-quadratic-additive equation, J. Math. Inequal., 12 (2018), 43–61.
    [16] S. M. S. M. Mosadegh, E. Movahednia, Stability of preserving lattice cubic functional equation in Menger probabilistic normed Riesz spaces, J. Fixed Point Theory Appl., 20 (2018), 1–15. doi: 10.1007/s11784-018-0489-6
    [17] M. S. Moslehian, Superstability of higher derivations in multi-Banach algebras, Tamsui Oxf. J. Math. Sci., 24 (2008), 417–427.
    [18] M. S. Moslehian, K. Nikodem, D. Popa, Asymptotic aspect of the quadratic functional equation in multi-normed spaces, J. Math. Anal. Appl., 355 (2009), 717–724. doi: 10.1016/j.jmaa.2009.02.017
    [19] E. Movahednia, Fuzzy stability of quadratic functional equations in general cases, ISRN Math. Anal., 10 (2011), 553–560.
    [20] E. Movahednia, S. M. S. M. Mosadegh, C. Park, D. Shin, Stability of a lattice preserving functional equation on riesz space: Fixed point alternative, J. Comput. Anal. Appl., 21 (2016), 83–89.
    [21] E. Movahednia, M. Mursaleen, Stability of a generalized quadratic functional equation in intuitionistic fuzzy 2-normed space, Filomat, 30 (2016), 449–457. doi: 10.2298/FIL1602449M
    [22] V. Radu, The fixed point alternative and the stability of functional equations, Fixed Point Theory, 4 (2003), 91–96.
    [23] Th. M. Rassias, On the stability of the linear mapping in Banach spaces, Proc. Am. Math. Soc., 72 (1978), 297–300. doi: 10.1090/S0002-9939-1978-0507327-1
    [24] K. Y. N. Sayar, A. Bergam, Approximate solutions of a quadratic functional equation in 2-Banach spaces using fixed point theorem, J. Fixed Point Theory Appl., 22 (2020), 1–16. doi: 10.1007/s11784-019-0746-3
    [25] S. M. Ulam, Problems in Modern Mathematics, Wiley, New York, 1964.
    [26] J. Vanterler da C. Sousa, E. Capelas de Oliveira, F. G. Rodrigues, Ulam-Hyers stabilities of fractional functional differential equations, AIMS Math., 5 (2020), 1346–1358. doi: 10.3934/math.2020092
  • This article has been cited by:

    1. Lin Chen, Xiaolin Luo, The stability of high ring homomorphisms and derivations on fuzzy Banach algebras, 2024, 22, 2391-5455, 10.1515/math-2024-0069
  • Reader Comments
  • © 2021 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(2600) PDF downloads(131) Cited by(1)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog