Abstract
In this paper we consider a generalization of d’Alembert’s equation and Wilson’s equation on commutative semigroups using only the semigroup operation, ie. we consider the functional equation
where \(f,h:S\rightarrow \mathbb {K}\), \((S,+)\) is a commutative semigroup, \(\mathbb {K}\) is a quadratically closed field, \(\text {char}\,\mathbb {K}\ne 2\).
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
If we look at d’Alembert’s functional equation
on a group S, then we have two possible ways of generalizing it to semigroups.
The first one is the functional equation
where \((S,+)\) is an abelian semigroup, \(\sigma \in \text {Aut}(S)\) is an involution, which has been studied by many mathematicians, (see [3, 4, 7, 12,13,14, 16, 18, 19]). For non-abelian groups the solutions of d’Alembert’s functional equation may be different from those of the abelian case (see [8, 10, 11, 20]).
The above equation has a generalization of the form
where \((G,+)\) is a locally compact group, K is a compact subgroup of the automorphism group of G with the normalized Haar measure \(\mu \), \(f:G\rightarrow \mathbb {C}\). It is a generalization of the cosine equation and it is studied in the theory of group representations, being the relation defining K-spherical functions (for the terminology see [5, p. 88]). This equation has been studied by many mathematicians (for example see [6, 15, 17, 21, 22]).
The second way is the functional equation
which we obtain by the substitution \(x\mapsto x+y\). This equation is equivalent to d’Alembert’s functional equation on groups and we will show that its solutions are the same as those of d’Alembert’s functional equation.
2 Preliminaries
Throughout the present paper, we assume that \((S,+)\) is an abelian semigroup and the relation \(\sim \subseteq S\times S\) is given by
\(\mathbb {K}\) is a quadratically closed field, \(\mathrm{char}\,\mathbb {K}\ne 2\).
Lemma 2.1
([15, Lemma 2.2]) The relation \(\sim \) given by (2.1) is an equivalence relation, \(S/_\sim \) with the operation \(+:{S/_\sim }^2 \rightarrow S/_\sim \) defined by
is a cancellative abelian semigroup and the function \(\varkappa :S\rightarrow S/_\sim \) given by
is a semigroup epimorphism.
Theorem 2.2
([18, Theorem 1]) Let \(\sigma :S\rightarrow S\) be an involution, \(f:S\rightarrow \mathbb {K}\). Then f satisfies Eq. (1.2) iff there exists an exponential function \(m:S\rightarrow \mathbb {K}\) such that
The exponential function m from the above theorem is on groups either zero everywhere or non-zero everywhere (e.g. \(m:S\rightarrow \mathbb {K}^*\) is a homomorphism). Similarly on semigroups, m has the same zero behavior as in groups iff f satisfies
(see [15, Theorem 2.8]). But generally on semigroups there may exist exponential functions which have zeros on some non-trivial subset of S.
Example
Let \(c\in \mathbb {K}\setminus \{0\}\), \(\mathbb {N}_0=\mathbb {N}\cup \{0\}\), \(S=\mathbb {N}_0\times \mathbb {N}_0\), \(\sigma :S\rightarrow S\) be a function given by \(\sigma (n,k)=(k,n)\) for \(n,k\in \mathbb {N}_0\). We define functions \(f,m:S\rightarrow \mathbb {K}\) by the formulas
Let further \(n,k,p,q\in \mathbb {N}_0\). We observe that:
-
if \(k\ge 1\) or \(q\ge 1\), then
$$\begin{aligned}&m(n+p,k+q)=0=m(n,k)\cdot m(p,q). \end{aligned}$$ -
if \(k=q=0\), then
$$\begin{aligned} m(n+p,k+q)&=m(n+p,0)=(2c)^{n+p}=(2c)^n \cdot (2c)^p\\&=m(n,0)\cdot m(p,0)=m(n,k)\cdot m(p,q). \end{aligned}$$
Hence m is an exponential function. We observe also that
-
if \(k=n=0\), then
$$\begin{aligned}&m(n,k)+m(k,n)=2=2f(n,k). \end{aligned}$$ -
if \(k,n\ge 1\), then
$$\begin{aligned}&m(n,k)+m(k,n)=0=f(n,k). \end{aligned}$$ -
if \(k\ge 1\) and \(n=0\), then
$$\begin{aligned}&m(n,k)+m(k,n)=0+(2c)^k=2\cdot 2^{k-1}c^k=f(n,k). \end{aligned}$$ -
if \(k=0\) and \(n\ge 1\), then
$$\begin{aligned}&m(n,k)+m(k,n)=(2c)^n+0=2\cdot 2^{n-1}c^n=f(n,k). \end{aligned}$$
Hence f satisfies Eq. (2.4) and in view of Theorem 2.2 it satisfies Eq. (1.2). We have also
3 Main results
Lemma 3.1
Let \(f:S\rightarrow \mathbb {K}\) satisfy Eq. (1.3). Then the function \(\widetilde{f}:S/_{\sim } \rightarrow \mathbb {K}\) given by the formula \(\widetilde{f}(\varkappa (x))=f(x)\) for \(x\in S\) is well-defined and \(\widetilde{f}\) satisfies Eq. (1.3).
Proof
Let \(x,y,z\in S\) be such that \(x+z=y+z\). Then
so \(\widetilde{f}\) is well-defined. We have also
\(\square \)
Lemma 3.2
Assume that S is abelian and cancellative, G is an abelian group such that \(G=S-S\). Let \(f:S\rightarrow \mathbb {K}\) satisfy Eq. (1.3). Then the function \(F:G\rightarrow \mathbb {K}\) given by
is well-defined, \(F|_S =f\) and F satisfies the equation
Proof
Let \(x,y,u,v\in S\), \(x-y=u-v\). Then \(x+v=y+u\) and
which means that
so F is well-defined. We observe also that
Now we show that
Indeed, for \(x,y,z\in S\) we have
Hence, for \(x,y,u,v\in S\) we get
which ends the proof.\(\square \)
Using Lemmas 2.1, 3.1, 3.2 and the fact that \(\varkappa \) is a homomorphism we easily obtain the following result.
Corollary 3.3
Let G be an abelian group such that \(G=S/_\sim -S/_\sim \).
-
1.
Let \(f:S\rightarrow \mathbb {K}\) satisfy Eq. (1.3). Then the function \(F:G\rightarrow \mathbb {K}\) given by
$$\begin{aligned} F(\varkappa (x)-\varkappa (y))=2f(x)f(y)-f(x+y),\ x,y\in S, \end{aligned}$$(3.3)is well-defined, \(F\circ \varkappa =f\) and F satisfies d’Alembert’s functional equation.
-
2.
Let \(F:G\rightarrow \mathbb {K}\) satisfy d’Alembert’s functional equation. Then \(f=F\circ \varkappa :S\rightarrow \mathbb {K}\) satisfies Eq. (1.3).
Theorem 3.4
Let \(f:S\rightarrow \mathbb {K}\) be a non-zero function. Then f satisfies Eq. (1.3) iff there exists a homomorphism \(m:S\rightarrow \mathbb {K}^*\) such that
Proof
It is easy to check that the function given by (3.4) satisfies Eq. (1.3).
Assume that f satisfies Eq. (1.3). In view of Corollary 3.3 there exists a function \(F:G\rightarrow \mathbb {K}\) such that \(F\circ \varkappa =f\), F satisfies d’Alembert’s functional equation, where G is an abelian group such that \(G=S/_\sim - S/_\sim \). Hence there exists a homomorphism \(M:G\rightarrow \mathbb {K}^*\) such that
We define \(m:S\rightarrow \mathbb {K}^*\) by the formula
Since M and \(\varkappa \) are homomorphisms, m is a homomorphism. We have also
which ends the proof.\(\square \)
A well-known generalization of d’Alembert’s functional equation is Wilson’s functional equation (see e.g. [1] for references).
Solutions of this equation can be found as a special case of some more general functional equation in [2], but we use a more readable result from the paper [9, Theorem 8].
Theorem 3.5
Let G be an abelian group, \(F,H:G\rightarrow \mathbb {K}\). The ordered pair (F, H) satisfies Wilson’s functional equation
iff F, H have one of the following forms:
-
1.
\(H=0\) and F is arbitrary;
-
2.
\(F(x)=\frac{M(x)+M(x)^{-1}}{2}\), \(H(x)=c\frac{M(x)+M(x)^{-1}}{2}+d\frac{M(x)-M(x)^{-1}}{2}\) for \(x\in G\);
-
3.
\(F(x)=M(x)\), \(H(x)=M(x)\Big ( A(x)+c \Big )\), \(M(x)\in \{1,-1\}\) for \(x\in G\);
where \(M:G\rightarrow \mathbb {K}^*\) is a homomorphism, \(A:G\rightarrow \mathbb {K}\) is additive, \(c,d\in \mathbb {K}\).
We can equivalently write Wilson’s functional equation in the form
and now we can consider it on semigroups.
Lemma 3.6
Let \(f,h:S\rightarrow \mathbb {K}\), \(h\ne 0\), (f, h) satisfies (3.5) and. Then f is a non-zero function which satisfies (1.3).
Proof
We observe that
Suppose that there exist \(y,z\in S\) such that \(h(x+y+2z)=0\) for all \(x\in S\). Then
so
and
which gives us a contradiction. Hence f satisfies (1.3).
Suppose that \(f=0\). Then
which means that \(h(2x)=0\) for all \(x\in S\). We have also
so
which gives us a contradiction.\(\square \)
Lemma 3.7
Let \(f,h:S\rightarrow \mathbb {K}\), \(h\ne 0\), (f, h) satisfies Eq. (3.5). Then functions \(\widetilde{f},\widetilde{h}:S/_{\sim } \rightarrow \mathbb {K}\) given by the formulas \(\widetilde{f}(\varkappa (x))=f(x)\), \(\widetilde{h}(\varkappa (x))=h(x)\) for \(x\in S\) are well-defined and \((\widetilde{f},\widetilde{h})\) satisfies Eq. (3.5).
Proof
In view of Lemmas 3.1 and 3.6 the map \(\widetilde{f}\) is well-defined.
Let \(x,y,z\in S\) be such that \(x+z=y+z\). Then
so \(\widetilde{h}\) is well-defined. We have also
\(\square \)
Lemma 3.8
Assume that S is abelian and cancellative, G is an abelian group such that \(G=S-S\). Let \(f,h:S\rightarrow \mathbb {K}\), \(h\ne 0\), (f, h) satisfies Eq. (3.5). Then functions \(F,H:G\rightarrow \mathbb {K}\) given by
are well-defined, \(F|_S =f\), \(H|_S =h\) and (F, H) satisfies the equation
Proof
In view of Lemmas 3.2 and 3.6 the map F is well-defined.
Let \(x,y,u,v\in S\), \(x-y=u-v\). Then \(x+v=y+u\) and
which means that
so H is well-defined. We observe also that
Now we show that
Indeed, we have
Hence, for \(x,y,u,v\in S\) we get
which ends the proof.\(\square \)
Using Lemmas 2.1, 3.7, 3.8 and the fact that \(\varkappa \) is a homomorphism we easily obtain the following result.
Corollary 3.9
Let G be an abelian group such that \(G=S/_\sim -S/_\sim \).
-
1.
Let \(f,h:S\rightarrow \mathbb {K}\), \(h\ne 0\), (f, h) satisfies Eq. (3.5). Then functions \(F,H: G\rightarrow \mathbb {K}\) given by
$$\begin{aligned}&F(\varkappa (x)-\varkappa (y))=2f(x)f(y)-f(x+y),\ x,y\in S,\end{aligned}$$(3.9)$$\begin{aligned}&H(\varkappa (x)-\varkappa (y))=2h(x)f(y)-h(x+y),\ x,y\in S, \end{aligned}$$(3.10)are well-defined, \(F\circ \varkappa =f\), \(H\circ \varkappa =h\) and (F, H) satisfies Wilson’s functional equation.
-
2.
Let \(F,H:G\rightarrow \mathbb {K}\), \(f=F\circ \varkappa , h=H\circ \varkappa :S\rightarrow \mathbb {K}\), (F, H) satisfies Wilson’s functional equation. Then (f, h) satisfies Eq. (3.5).
Theorem 3.10
Let \(f,h:S\rightarrow \mathbb {K}\). Then (f, h) satisfies (3.5) iff f, h have one of the following forms:
-
1.
\(h=0\) and f is arbitrary;
-
2.
\(f(x)=\frac{m(x)+m(x)^{-1}}{2}\), \(h(x)=c\frac{m(x)+m(x)^{-1}}{2}+d\frac{m(x)-m(x)^{-1}}{2}\) for \(x\in S\);
-
3.
\(f(x)=m(x)\), \(h(x)=m(x)\Big ( a(x)+c \Big )\), \(m(x)\in \{1,-1\}\) for \(x\in S\);
where \(m:S\rightarrow \mathbb {K}^*\) is a homomorphism, \(a:S\rightarrow \mathbb {K}\) is additive, \(c,d\in \mathbb {K}\).
Proof
It is easy to check that for functions f, h given by the forms 1–3 the pair (f, h) satisfies Eq. (3.5).
Assume that (f, h) satisfies Eq. (3.5). In view of Corollary 3.9 there exist functions \(F,H:G\rightarrow \mathbb {K}\) such that \(F\circ \varkappa =f\), \(H\circ \varkappa =h\), (F, H) satisfies Wilson’s functional equation, where G is an abelian group such that \(G=S/_\sim - S/_\sim \). Hence we get that F, H have one of the following forms:
-
1.
\(H=0\) and F is arbitrary;
-
2.
\(F(x)=\frac{M(x)+M(x)^{-1}}{2}\), \(H(x)=c\frac{M(x)+M(x)^{-1}}{2}+d\frac{M(x)-M(x)^{-1}}{2}\) for \(x\in G\);
-
3.
\(F(x)=M(x)\), \(H(x)=M(x)\Big ( A(x)+c \Big )\), \(M(x)\in \{1,-1\}\) for \(x\in G\);
where \(M:G\rightarrow \mathbb {K}^*\) is a homomorphism, \(A:G\rightarrow \mathbb {K}\) is additive, \(c,d\in \mathbb {K}\).
We define \(m:S\rightarrow \mathbb {K}^*\), \(a:S\rightarrow \mathbb {K}\) by the formulas
Since M and \(\varkappa \) are homomorphisms, m is a homomorphism. Since A and \(\varkappa \) are additive, a is additive. We have also:
-
In case 2
$$\begin{aligned} f(x)&=F(\varkappa (x))=\frac{M(\varkappa (x))+M(\varkappa (x))^{-1}}{2}=\frac{m(x)+m(x)^{-1}}{2},\ x\in S,\\ h(x)&=H(\varkappa (x))= c\frac{M(\varkappa (x))+M(\varkappa (x))^{-1}}{2}+d\frac{M(\varkappa (x))-M(\varkappa (x))^{-1}}{2}\\&=c\frac{m(x)+m(x)^{-1}}{2}+d\frac{m(x)-m(x)^{-1}}{2},\ x\in S, \end{aligned}$$ -
In case 3
$$\begin{aligned}&f(x)=F(\varkappa (x))=M(\varkappa (x))=m(x),\ x\in S,\\&h(x)=H(\varkappa (x))=M(\varkappa (x))\Big ( A(\varkappa (x))+c \Big )=m(x)\Big ( a(x)+c \Big ),\ x\in S, \end{aligned}$$
which ends the proof.\(\square \)
References
Aczél, J., Dhombres, J.: Functional Equations in Several Variables. Cambridge University Press, Cambridge (1989)
Aczél, J., Chung, J.K., Ng, C.T.: Symmetric second differences in product form on groups. Top. Math. Anal. Ser. Pure Math. 11, 1–12 (1989)
Badora, R.: On the d’Alembert type functional equation in Hilbert algebras. Funkcial. Ekvac. 43, 405–418 (2000)
Baker, J.A.: D’Alembert’s functional equation in Banach algebras. Acta Sci. Math. (Szeged) 32, 225–234 (1971)
Benson, C., Jenkins, J., Ratcliff, G.: On Gelfand pairs associated with solvable Lie groups. Trans. Am. Math. Soc. 321, 85–116 (1990)
Chojnacki, W.: On some functional equation generalizing Cauchy’s and d’Alembert’s functional equations. Colloq. Math. 55, 169–178 (1988)
Corovei, I.: The cosine functional equation for nilpotent groups. Aequ. Math. 15, 99–106 (1977)
Corovei, I.: The d’Alembert functional equation on metabelian groups. Aequ. Math. 57, 201–205 (1999)
Corovei, I.: Wilson’s functional equation on \(P_3\)-groups. Aequ. Math. 61, 212–220 (2001)
Davison, T.M.K.: D’Alembert’s functional equation on topological groups. Aequ. Math. 76(1–2), 33–53 (2008)
Davison, T.M.K.: D’Alembert’s functional equation on topological monoids. Publ. Math. Debr. 75(1/2), 41–66 (2009)
Förg-Rob, W., Schwaiger, J.: A generalization of the cosine equation to \(n\) summands. Grazer Math. Ber. 316, 219–226 (1992)
Gajda, Z.: A remark on the talk of W. Förg-Rob. Grazer Math. Ber. 316, 234–237 (1992)
Kannappan, P.: The functional equation \(f(xy)+f(xy^{-1})=2f(x)f(y)\) for groups. Proc. Am. Math. Soc. 19, 69–74 (1968)
Łukasik, R.: \(K\)-spherical functions on abelian semigroups. Bull. Aust. Math. Soc. 96, 479–486 (2017). https://doi.org/10.1017/S0004972717000417
Penney, R.C., Rukhin, A.L.: D’Alembert’s functional equation on groups. Proc. Am. Math. Soc. 77, 73–80 (1979)
Shin’ya, H.: Spherical matrix functions and Banach representability for locally compact motion groups. Jpn. J. Math. 28, 163–201 (2002)
Sinopoulos, P.: Functional equations on semigroups. Aequ. Math. 59, 255–261 (2000)
Stetkær, H.: D’Alembert’s equation and spherical functions. Aequ. Math. 48, 220–227 (1994)
Stetkær, H.: D’Alembert’s functional equations on metabelian groups. Aequ. Math. 56(3), 306–320 (2000)
Stetkær, H.: Functional equations and matrix-valued spherical functions. Aequ. Math. 69, 271–292 (2005)
Stetkær, H.: On operator-valued spherical functions. J. Funct. Anal. 224, 338–351 (2005)
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Łukasik, R. D’Alembert’s and Wilson’s equations on semigroups. Aequat. Math. 94, 1269–1279 (2020). https://doi.org/10.1007/s00010-020-00708-3
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00010-020-00708-3