You are on page 1of 15

DOI 10.

1007/s10958-022-05811-1
Journal of Mathematical Sciences, Vol. 262, No. 2, April, 2022

Hausdorff analytic functions in a three-dimensional associative


noncommutative algebra
Vitalii Shpakivskyi, Tetiana Kuzmenko

(Presented by V.I. Ryazanov)

Dedicated to the 80th anniversary of the Corresponding Member of NAS of Ukraine


V.Ya. Gutlyanskii

Abstract. A class of H-analytic (differentiable by Hausdorff) functions in a three-dimensional noncommu-


 2 over the field C is considered. All H-analytic functions are described in the explicit form.
tative algebra A
The obtained description is applied to the integral representation of these functions, and the mentioned
functions are also applied when solving some PDEs.

Keywords. Noncommutative algebra, H-analytic function, constructive description, integral representa-


tion, Laplace equation, biharmonic equation.

1. Introduction

F. Hausdorff [1] has proposed a definition of an analytic function in a complex associative (commu-
tative or noncommutative) algebra A with the unit.Let e1 , e2 , . . . , en be a basis of A, let η1 , η2 , . . . , ηn
be the respective complex variables, and let η := nk=1 ηk ek be the hypercomplex variable. Consider
an A-valued function
n

f (η) = fk (η1 , . . . , ηn )ek ,
k=1
where the components fk are holomorphic functions of η1 , η2 , . . . , ηn in the sense of several complex
variables theory. Its complex differential is defined in the usual way as
n
 ∂fk
df (dη1 , dη2 , . . . , dηn ) := (η1 , η2 , . . . , ηn )dηj ek .
∂ηj
j,k=1

A function f is called H-analytic (or Hausdorff-analytic) function of the hypercomplex


 variable η
if the differential df is a C-linear homogeneous polynomial of the differential dη := nk=1 dηk ek , i.e.,
n 2

df = As (η) dη Bs (η) ,
s=1

where As and Bs are some A-valued functions of the variable η.


 (η) :=
n 2
For an H-analytic function f , the element fH s=1 As Bs is called the H-derivative (Haus-
dorff derivative) of the function f (η) at the point η.
Translated from Ukrains’kiı̆ Matematychnyı̆ Visnyk, Vol. 19, No. 1, pp. 103–120, January–March, 2022.
Original article submitted December 20, 2021
This research is partially supported by budget program “Support for the development of priority areas of research”
(KPKVK 6541230).

1072 – 3374/22/2622–0207 
c 2022 Springer Science+Business Media, LLC 207
There exists a kind of a “real version” of Hausdorff’s approach (see, e.g, [2–4]), where A is a real
algebra, η1 , η2 , . . . , ηn are real variables, the components fk are R-valued functions of η, and all fk are
assumed to be R-differentiable at a given point of Rn . Following this way, H-analytic functions in the
algebra of real quaternions H were considered in the paper [5].
In the paper [6], this approach was realized in the special three-dimensional noncommutative algebra
 2 over the field of real numbers R. In the cited paper, all H-analytic functions with values in the
A
algebra A 2 are explicitly described.
In this paper, we consider a class of H-analytic functions in the three-dimensional noncommutative
algebra A 2 over the field of complex numbers C with the basis {I1 , I2 , ρ}, whose elements satisfy the
following multiplication rules:
· I 1 I2 ρ
I1 I 1 0 0
.
I2 0 I2 ρ
ρ ρ 0 0

Note that the subalgebra with the basis {I1 , I2 } is the algebra of bicomplex numbers B(C) or Segre’s
algebra of commutative quaternions (see, e.g., [7, 8]).
In the paper [9], G-monogenic (locally bounded and differentiable by Gâteaux) mappings defined
in the domains of the three-dimensional subspace of the algebra B(C) and taking values in the algebra
 2 are considered. Such mappings are described by means of three holomorphic functions of a complex
A
variable.
In the paper [10], a class of the differentiable functions defined by the equalities

dΦ = dζ · Φ (ζ) and dΦ = Φ (ζ) · dζ

 2 . All these functions were explicitly constructed and the obtained


were also considered in the algebra A
descriptions were applied to the integral representation of these functions and their representation by
series.

2. H-analytic functions

We now apply the Hausdorff approach to the functions with values in the algebra A  2.
 
Let Ω = {ζ = ξ1 I1 + ξ2 I2 + ξ3 ρ ∈ A2 } be a domain in the algebra A2 . Further, we will identify the
domain Ω with the domain ΩC = {(ξ1 , ξ2 , ξ3 ) ∈ C3 }.
Consider a function Φ : Ω → A  2 of the variable ζ = ξ1 I1 + ξ2 I2 + ξ3 ρ given as

Φ(ζ) = f1 (ξ1 , ξ2 , ξ3 )I1 + f2 (ξ1 , ξ2 , ξ3 )I2 + f3 (ξ1 , ξ2 , ξ3 )ρ, (2.1)

where ξ1 , ξ2 , ξ3 are independent complex variables and f1 , f2 , f3 are holomorphic functions of three
complex variables.
A function Φ : Ω → A  2 of form (2.1) is called H-analytic in the domain Ω, if Φ is differentiable by
Hausdorff at every point ζ ∈ Ω, i.e., if the differential dΦ is a C-linear homogeneous polynomial of the
differential dζ = dξ1 I1 + dξ2 I2 + dξ3 ρ, i.e.,

9

dΦ = As (ζ) dζ Bs (ζ) , (2.2)
s=1

208
where As and Bs are some A  2 -valued functions of the variable ζ, and, moreover, the second-order
2
differential d Φ is a homogeneous polynomial of dζ, i.e.,
9

d2 Φ = Âs (ζ) dζ B̂s (ζ) dζ Ĉs (ζ), (2.3)
s=1

 2 -valued functions of the variable ζ.


where Âs , B̂s , Ĉs are some A
In this case, for any ζ ∈ Ω, the element
9

ΦH (ζ) := As Bs (2.4)
s=1

is called the H-derivative of the function Φ at the point ζ.


Theorem 2.1. A function Φ : Ω → A  2 of form (2.1), where fk : Ω → C3 are holomorphic functions,
 2 if and only if the conditions
is H-analytic in the domain Ω ⊂ A
∂f1 ∂f1 ∂f2 ∂f2
= = = = 0, (2.5)
∂ξ2 ∂ξ3 ∂ξ1 ∂ξ3

∂ 2 f3
=0 (2.6)
∂ξ32
are satisfied.
Proof. Necessity. Let

As = as1 I1 + as2 I2 + as3 ρ, Bs = bs1 I1 + bs2 I2 + bs3 ρ

for s = 1, 2, . . . , 9. Then, due to the H-analyticity of the function Φ, we have the equality
 
∂f1 ∂f1 ∂f1
dξ1 + dξ2 + dξ3 I1
∂ξ1 ∂ξ2 ∂ξ3
 
∂f2 ∂f2 ∂f2
+ dξ1 + dξ2 + dξ3 I2
∂ξ1 ∂ξ2 ∂ξ3
 
∂f3 ∂f3 ∂f3
+ dξ1 + dξ2 + dξ3 ρ
∂ξ1 ∂ξ2 ∂ξ3
9

= (as1 I1 + as2 I2 + as3 ρ)(dξ1 I1 + dξ2 I2 + dξ3 ρ)(bs1 I1 + bs2 I2 + bs3 ρ)
s=1
9

= as1 bs1 dξ1 I1 + as2 bs2 dξ2 I2 + (as2 bs3 dξ2 + as2 bs1 dξ3 + as3 bs1 dξ1 )ρ.
s=1
Since
 9
∂f1 ∂f1 ∂f1
dξ1 + dξ2 + dξ3 = as1 bs1 dξ1 ,
∂ξ1 ∂ξ2 ∂ξ3
s=1
then
∂f1 ∂f1
= = 0.
∂ξ2 ∂ξ3

209
Similarly,
 9
∂f2 ∂f2 ∂f2
dξ1 + dξ2 + dξ3 = as2 bs2 dξ2 .
∂ξ1 ∂ξ2 ∂ξ3
s=1

Therefore,
∂f2 ∂f2
= = 0.
∂ξ1 ∂ξ3
Now let
Âs = âs1 I1 + âs2 I2 + âs3 ρ,
B̂s = b̂s1 I1 + b̂s2 I2 + b̂s3 ρ,
Ĉs = ĉs1 I1 + ĉs2 I2 + ĉs3 ρ
for s = 1, 2, . . . , 9. Taking into account conditions (2.3) and (2.5), we have

∂ 2f ∂ 2f ∂ 2 f3 2 ∂ 2 f3 2 ∂ 2 f3 2
1 2
d2 Φ = dξ 2
I 1 + dξ 2
I 2 + dξ + dξ + dξ
∂ξ12 1 ∂ξ22 2 ∂ξ12 1 ∂ξ22 2 ∂ξ32 3

∂ 2 f3 ∂ 2 f3 ∂ 2 f3
+2 dξ1 dξ2 + 2 dξ1 dξ3 + 2 dξ2 dξ3 ρ
∂ξ1 ∂ξ2 ∂ξ1 ∂ξ3 ∂ξ2 ∂ξ3
9

= (âs1 I1 + âs2 I2 + âs3 ρ)(dξ1 I1 + dξ2 I2 + dξ3 ρ)(b̂s1 I1 + b̂s2 I2 + b̂s3 ρ)×
s=1

×(dξ1 I1 + dξ2 I2 + dξ3 ρ)(ĉs1 I1 + ĉs2 I2 + ĉs3 ρ)


9

= âs1 b̂s1 ĉs1 dξ12 I1 + âs2 b̂s2 ĉs2 dξ22 I2
s=1

+ âs3 b̂s1 ĉs1 dξ12 + âs2 b̂s2 ĉs3 dξ22 + âs2 b̂s3 ĉs1 dξ1 dξ2

+âs2 b̂s1 ĉs1 dξ1 dξ3 + âs2 b̂s2 ĉs1 dξ2 dξ3 ρ.

Equating the corresponding components in the left- and right-hand sides of this equality, we obtain
equality (2.6).
Sufficiency. Using the holomorphicity of functions f1 , f2 , f3 , conditions (2.5), and the multiplica-
 2 , we have
tion table of the algebra A
 
∂f1 ∂f1 ∂f1
dΦ = df1 I1 + df2 I2 + df3 ρ = dξ1 + dξ2 + dξ3 I1
∂ξ1 ∂ξ2 ∂ξ3
   
∂f2 ∂f2 ∂f2 ∂f3 ∂f3 ∂f3
+ dξ1 + dξ2 + dξ3 I2 + dξ1 + dξ2 + dξ3 ρ
∂ξ1 ∂ξ2 ∂ξ3 ∂ξ1 ∂ξ2 ∂ξ3
 
∂f1 ∂f2 ∂f3 ∂f3 ∂f3
= dξ1 I1 + dξ2 I2 + dξ1 + dξ2 + dξ3 ρ
∂ξ1 ∂ξ2 ∂ξ1 ∂ξ2 ∂ξ3
∂f1 ∂f2 ∂f3 ∂f3 ∂f3
= I1 dζ I1 + I2 dζ I2 + ρ dζ I1 + I2 dζ ρ + I2 dζ I1 . (2.7)
∂ξ1 ∂ξ2 ∂ξ1 ∂ξ2 ∂ξ3

210
Now using conditions (2.5) and (2.6), we show that the second differential d2 Φ is of form (2.3):

∂ 2f ∂ 2f ∂ 2 f 3 2 ∂ 2 f3 2
1 2
d2 Φ = dξ 2
1 I 1 + dξ 2
2 I 2 + dξ + dξ
∂ξ12 ∂ξ22 ∂ξ12 1 ∂ξ22 2

∂ 2 f3 ∂ 2 f3 ∂ 2 f3
+2 dξ1 dξ2 + 2 dξ1 dξ3 + 2 dξ2 dξ3 ρ
∂ξ1 ∂ξ2 ∂ξ1 ∂ξ3 ∂ξ2 ∂ξ3
∂ 2 f1 ∂ 2 f2
= 2 I1 dζ I1 dζ I1 + I2 dζ I2 dζ I2
∂ξ1 ∂ξ22
∂ 2 f3 ∂ 2 f3
+ ρ dζ I 1 dζ I 1 + I2 dζ I2 dζ ρ
∂ξ12 ∂ξ22
∂ 2 f3 ∂ 2 f3 ∂ 2 f3
+2 I2 dζ ρ dζ I1 + 2 I2 dζ I1 dζ I1 + 2 I2 dζ I2 dζ I1 ,
∂ξ1 ∂ξ2 ∂ξ1 ∂ξ3 ∂ξ2 ∂ξ3
which means the H-analyticity of the function Φ.

Theorem 2.2. If the function Φ : Ω → A  2 is H-analytic in the domain Ω, then its derivative Φ (ζ)
H
exists and it is independent of the choice of the functions As , Bs in equality (2.2). In this case,
 
 ∂f1 ∂f2 ∂f3 ∂f3
ΦH (ζ) = I1 + I2 + + ρ. (2.8)
∂ξ1 ∂ξ2 ∂ξ1 ∂ξ2
Proof. In Theorem 2.1, it is proved that, due to the H-analyticity of the function Φ, the differential
dΦ exists and it is of form (2.7). Therefore, using equality (2.4), we obtain equality (2.8).

3. Constructive description and integral representation of H-analytic functions

Due to equalities (2.5) and taking into account that the functions f1 , f2 are holomorphic, we have

f1 (ξ1 , ξ2 , ξ3 ) = f1 (ξ1 ), f2 (ξ1 , ξ2 , ξ3 ) = f2 (ξ2 ).

Now integrating equality (2.6) twice, we have


∂f3
= ϕ1 (ξ1 , ξ2 ),
∂ξ3
f3 (ξ1 , ξ2 , ξ3 ) = ϕ1 (ξ1 , ξ2 )ξ3 + ϕ2 (ξ1 , ξ2 ),
where ϕ1 , ϕ2 are holomorphic functions of two complex variables.
Therefore, every H-analytic function can be expressed in the form

Φ(ζ) = f1 (ξ1 )I1 + f2 (ξ2 )I2 + [ϕ1 (ξ1 , ξ2 )ξ3 + ϕ2 (ξ1 , ξ2 )] ρ. (3.1)

In this paper, we will first obtain a description of an analytic function with values in algebra
using holomorphic functions of several complex variables. Previously, in particular, in commutative
algebras [8, 12, 17, 18], a constructive description of analytic functions was obtained using holomorphic
functions of one complex variable.
Let Ω3 := {(ξ1 , ξ2 ) ∈ C2 such that ∃ ξ3 ∈ C : (ξ1 , ξ2 , ξ3 ) ∈ Ω}. The following statement follows
immediately from equality (3.1) because its right-hand side is an H-analytic function in the domain
 2 : (ξ1 , ξ2 ) ∈ Ω3 } ≡ Ω3 × C.
Π := {ζ = ξ1 I1 + ξ2 I2 + ξ3 ρ ∈ A (3.2)

211
 2 can be extended to the H-analytic function in
Theorem 3.1. Every H-analytic function Φ : Ω → A
the domain Π.

Henceforth we will consider functions, which are H-analytic in domains of the form D × C, where
D ∈ C2 .
Note that using representation (3.1), we have
∂Φ ∂f1 ∂f3 ∂Φ ∂f2 ∂f3
= I1 + ρ, = I2 + ρ.
∂ξ1 ∂ξ1 ∂ξ1 ∂ξ2 ∂ξ2 ∂ξ2
Then formula (2.8) can be rewritten in the compact form
∂Φ ∂Φ
ΦH (ζ) = + . (3.3)
∂ξ1 ∂ξ2
Example 3.1. Let Φ(ζ) = ζ n for n ∈ N . If n = 1, then Φ1 (ζ) = ζ = ξ1 I1 + ξ2 I2 + ξ3 ρ and, using
formula (3.3), we have
∂Φ1 ∂Φ1
Φ1,H (ζ) = + = I1 + I2 = 1.
∂ξ1 ∂ξ2
If n = 2, then Φ2 (ζ) = ζ 2 = ξ12 I1 + ξ22 I2 + ξ3 (ξ1 + ξ2 )ρ and
∂Φ2 ∂Φ2
Φ2,H (ζ) = + = 2ξ1 I1 + ξ3 ρ + 2ξ2 I2 + ξ3 ρ
∂ξ1 ∂ξ2
= 2ξ1 I1 + 2ξ2 I2 + 2ξ3 ρ = 2ζ.
In the same way, the H-derivative for the function

Φ3 (ζ) = ζ 3 = ξ13 + ξ23 I2 + ξ3 ξ12 + ξ1 ξ2 + ξ22 ρ

is
∂Φ3 ∂Φ3
Φ3,H (ζ) = + = 3ξ12 I1 + ξ3 (2ξ1 + ξ2 )ρ + 2ξ2 I2 + ξ3 (ξ1 + 2ξ2 )ρ
∂ξ1 ∂ξ2
= 3ξ12 I1 + 3ξ22 I2 + 3ξ3 (ξ1 + ξ2 )ρ = 3ζ 2 .
By induction,
n−1

Φn (ζ) = ζ n = ξ1n I1 + ξ2n I2 + ξ3 ξ1n−k−1 ξ2k ρ
k=0
and
n−2

Φn
n,H )

= nξ1n−1 I1 + ξ3 (n − k − 1)ξ1n−k−2 ξ2k ρ
k=0
n−2

+nξ2n−1 I2 + ξ3 (n − k − 1)ξ1k ξ2n−k−2 ρ
k=0
 n−2


=n ξ1n−1 I1 + ξ2n−1 I2 + ξ3 ξ1n−k−2 ξ2k ρ = nζ n−1 .
k=0

 2 forms
Remark 3.1. It is easy to see that the set of H-analytic functions with values in the algebra A
a functional algebra since the product of two H-analytic functions is also an H-analytic function.

212
Integral representations of the differentiable, in some sense, functions are important for applica-
tions, especially for solving boundary value problems. For example, in the papers [12, 13], integral
representation of monogenic functions in the two-dimensional biharmonic algebra were applied to solv-
ing boundary value problems for the two-dimensional biharmonic equation. In the monograph [14],
integral representations of the axial-symmetric potential and the Stokes flow functions were obtained,
and such representations were applied to solve the Dirichlet boundary problem for the mentioned
potentials. Furthermore, in the papers [15, 19], integral representations of monogenic functions in a
commutative algebra were applied to construct exact solutions of the linear homogeneous PDEs with
constant coefficients.
It is easy to show that the inverse element to the element ζ = ξ1 I1 + ξ2 I2 + ξ3 ρ is of the form

1 1 ξ3
ζ −1 = I1 + I 2 − ρ, (3.4)
ξ1 ξ2 ξ1 ξ 2

where ξ1 = 0 and ξ2 = 0.
Let τ := t1 I1 + t2 I2 , where t1 , t2 ∈ C and dτ := dt1 dt2 . Then, the next equality follows from
equality (3.4):
1 1 ξ3
(τ − ζ)−1 = I1 + I2 − ρ. (3.5)
t1 − ξ 1 t2 − ξ 2 (t1 − ξ1 )(t2 − ξ2 )

Let D ⊂ C2 be an arbitrary domain on the plane (t1 , t2 ). Let

D1 := {t1 ∈ C such that ∃ t2 ∈ C : (t1 , t2 ) ∈ D},

D2 := {t2 ∈ C such that ∃ t1 ∈ C : (t1 , t2 ) ∈ D}.

In the domain D, we take an arbitrary subdomain G such that G = G1 × G2 , Gs is a domain on


a plane ts , s = 1, 2, with a piecewise smooth boundary ∂Gs , and ξs ∈ Gs , s = 1, 2. In what follows,
Γ := ∂G1 × ∂G2 .
Now we construct a representation of H-analytic functions as the sum of integrals of the form

1
f (t1 , t2 ) (τ − ζ)−1 dτ,
(2πi)2
Γ

where f (t1 , t2 ) is some holomorphic function of two complex variables in D.

 2 in a domain D × C can be expressed by


Theorem 3.2. Every H-analytic function Φ : D × C → A
the formula
 
1 f1 (t1 ) ϕ2 (t1 , t2 )
Φ(ζ) = I1 − ϕ1 (t1 , t2 )I2 + ρ (τ − ζ)−1 dτ
(2πi)2 t2 − ξ 2 t2 − ξ 2
Γ


1 f2 (t2 )
+ (τ − ζ)−1 dτ I2 , (3.6)
(2πi)2 t1 − ξ 1
Γ

where fs is some holomorphic function in the domain Ds , s = 1, 2, and ϕ1 , ϕ2 are some holomorphic
functions in the domain D.

213
Proof. Let us show that representation (3.6) can be reduced to (3.1) by equivalent transformations.
Using equality (3.5) and the multiplication table of the algebra, equality (3.6) takes the form

1 f1 (t1 ) ϕ1 (t1 , t2 ) ϕ1 (t1 , t2 )ξ3
2
I1 − I2 + ρ+
(2πi) (t1 − ξ1 )(t2 − ξ2 ) t2 − ξ 2 (t1 − ξ1 )(t2 − ξ2 )
Γ
 5
ϕ2 (t1 , t2 ) f2 (t2 ) 
+ ρ+ I2 dτ =: Jk .
(t1 − ξ1 )(t2 − ξ2 ) (t1 − ξ1 )(t2 − ξ2 )
k=1

Now, applying the Cauchy formula to functions of several complex variables (see, e.g., [16, p. 45]), we
have
J1 = f1 (ξ1 )I1 , J2 = 0, J3 = ϕ1 (ξ1 , ξ2 )ξ3 ρ,
J4 = ϕ2 (ξ1 , ξ2 )ρ, J5 = f2 (ξ2 )I2 .
Thus, the function Φ is of form (3.1).

4. Application of H-analytic functions to solve linear PDEs

4.1 Solutions of three-dimensional Laplace equation


Consider the three-dimensional Laplace equation

∂2Φ ∂2Φ ∂2Φ


+ + = 0. (4.1)
∂x2 ∂y 2 ∂z 2

 2 have the following decompositions in the algebra A


Let vectors e1 , e2 , e3 of A  2:

e1 = k1 I1 + k2 I2 + k3 ρ,
e2 = m1 I1 + m2 I2 + m3 ρ,
e3 = n1 I1 + n2 I2 + n3 ρ,

where ks , ms , ns ∈ C, s = 1, 2, 3, and ζ := xe1 + ye2 + ze3 for x, y, z ∈ R.


In this notation, the variable ζ takes the form

ζ = (k1 x + m1 y + n1 z) I1 + (k2 x + m2 y + n2 z) I2 + (k3 x + m3 y + n3 z) ρ,

i.e., instead of the independent complex components ξ1 , ξ2 , ξ3 , we have the components of the special
form
ξs := ks x + ms y + ns z, (4.2)
where ks , ms , ns ∈ C, s = 1, 2, 3.
It is obvious that the components f1 (ξ1 ) and f2 (ξ2 ) of the H-analytic function (3.1) satisfy the
Laplace equation as holomorphic functions, if the following equations are valid:

k12 + m21 + n21 = 0, k22 + m22 + n22 = 0. (4.3)

Now consider the third component of the H-analytic function

f3 (ξ1 , ξ2 , ξ3 ) = ϕ1 (ξ1 , ξ2 )ξ3 + ϕ2 (ξ1 , ξ2 ) (4.4)

214
and make the substitution
∂ 2 f3 ∂ 2 f3 ∂ 2 f3
+ +
∂x2 ∂y 2 ∂z 2
 2ϕ

∂ 1 ∂ 2 ϕ1 ∂ 2 ϕ1
= ξ3 k12 2 + 2k1 k2 + k22
∂ξ1 ∂ξ1 ∂ξ2 ∂ξ22
∂ϕ2 ∂ϕ2 ∂ 2 ϕ2 ∂ 2 ϕ2 2
2 ∂ ϕ2
+2k1 k3 + 2k2 k3 + k12 + 2k 1 k 2 + k 2
∂ξ1 ∂ξ2 ∂ξ12 ∂ξ1 ∂ξ2 ∂ξ22
 
∂ 2 ϕ1 ∂ 2 ϕ1 2
2 ∂ ϕ1
+ξ3 m21 + 2m m
1 2 + m 2
∂ξ12 ∂ξ1 ∂ξ2 ∂ξ22
∂ϕ2 ∂ϕ2 ∂ 2 ϕ2 ∂ 2 ϕ2 2
2 ∂ ϕ2
+2m1 m3 + 2m2 m3 + m21 + 2m m
1 2 + m 2
∂ξ1 ∂ξ2 ∂ξ12 ∂ξ1 ∂ξ2 ∂ξ22
 2

2 ∂ ϕ1 ∂ 2 ϕ1 2
2 ∂ ϕ1
+ξ3 n1 + 2n1 n2 + n2
∂ξ12 ∂ξ1 ∂ξ2 ∂ξ22
∂ϕ2 ∂ϕ2 ∂ 2 ϕ2 ∂ 2 ϕ2 2
2 ∂ ϕ2
+2n1 n3 + 2n2 n3 + n21 + 2n 1 n 2 + n 2
∂ξ1 ∂ξ2 ∂ξ12 ∂ξ1 ∂ξ2 ∂ξ22
   
2 2 2
∂ 2 ϕ1 ∂ 2 ϕ2 2 2 2
∂ 2 ϕ1 ∂ 2 ϕ2
= k1 + m1 + n1 ξ3 + + k 2 + m2 + n 2 ξ3 +
∂ξ12 ∂ξ12 ∂ξ22 ∂ξ22
 
∂ 2 ϕ1 ∂ 2 ϕ2
+2 (k1 k2 + m1 m2 + n1 n2 ) ξ3 +
∂ξ1 ∂ξ2 ∂ξ1 ∂ξ2
∂ϕ2 ∂ϕ2
+2(k1 k3 + m1 m3 + n1 n3 ) + 2(k2 k3 + m2 m3 + n2 n3 ) .
∂ξ1 ∂ξ2
For the component f3 (ξ1 , ξ2 , ξ3 ) to satisfy equation (4.1), it is sufficient that the numbers ks , s = 1, 2, 3,
satisfy the following system: ⎧ 2

⎪ k1 + m21 + n21 = 0,


⎪ k 2 + m2 + n2 = 0,

⎨ 2 2 2
k1 k2 + m1 m2 + n1 n2 = 0,



⎪ k1 k3 + m1 m3 + n1 n3 = 0,



k2 k3 + m2 m3 + n2 n3 = 0,
which contains equations (4.3). The solutions of the obtained system are

k1 = −k2 , m1 = −m2 ,

k 1 k 3 + m1 m3
n1 = −n2 = ±i k12 + m21 , n3 = ±i  2 , (4.5)
k1 + m21
∀ k1 , k3 , m1 , m3 ∈ C,
where the both signs ± are simultaneously either + or −.
Therefore, equality (4.2) implies that the variables ξ1 , ξ2 , ξ3 are of the forms

ξ1 = k1 x + m1 y ± i k12 + m21 z, ξ2 = −ξ1 ,
(4.6)

215
k 1 k 3 + m1 m3
ξ3 = k3 x + m3 y ± i  2 ,
k1 + m21
where k1 , k3 , m1 , m3 are arbitrary complex numbers.
Thus, every H-analytic function Φ(ζ) satisfy equation (4.1) of the variables ξ1 , ξ2 , ξ3 of form (4.6).
Using representation (3.1), we can write an exact solution of equation (4.1) in the form of compo-
nents of an H-analytic function in A  2 . We will have
  
2 2
U 1 = U 2 = f 1 k 1 x + m1 y ± i k 1 + m1 z ,
  
k 1 k 3 + m1 m3
2 2
U 3 = k 3 x + m3 y ± i  ϕ 1 k 1 x + m1 y ± i k 1 + m1 z
k12 + m21
  
2
+ϕ2 k1 x + m1 y ± i k1 + m1 z , 2

where k1 , m1 , k2 , m2 are arbitrary complex numbers, and f1 , ϕ1 , ϕ2 are arbitrary holomorphic func-
tions of the corresponding complex variable.
Remark 4.1. Note that the obtained solutions of equation (4.1) exactly coincide with the solutions
obtained in the paper [15].

4.2 Solutions of biharmonic equation


Consider the two-dimensional biharmonic equation
∂4Φ ∂4Φ ∂4Φ
Δ2 Φ(x, y) := + 2 + = 0. (4.7)
∂x4 ∂x2 ∂y 2 ∂y 4
 2 have the following decompositions in the algebra A
Let the vectors e1 , e2 of A  2:

e1 = m1 I1 + m2 I2 + m3 ρ, e 2 = n1 I 1 + n2 I 2 + n3 ρ

and
ζ := xe1 + ye2 = (m1 x + n1 y) I1 + (m2 x + n2 y) I2 + (m3 x + n3 y) ρ,
where ms , ns ∈ C, s = 1, 2, 3.
It is easy to see that the components f1 (ξ1 ) and f2 (ξ2 ) of the H-analytic function (3.1) satisfy the
biharmonic equation as holomorphic functions if the following equalities are true:

(m21 + n21 )2 = 0, (m22 + n22 )2 = 0.

Now consider the third component of the H-analytic function

f3 (ξ1 , ξ2 , ξ3 ) = ϕ1 (ξ1 , ξ2 )ξ3 + ϕ2 (ξ1 , ξ2 ).

We make substitutions such that the function ϕ̃(ξ1 , ξ2 , ξ3 ) := ϕ1 (ξ1 , ξ2 )ξ3 satisfies the biharmonic
equation (4.7):
∂ 4 ϕ̃ ∂ 4 ϕ̃ ∂ 4 ϕ̃
+ 2 +
∂x4 ∂x2 ∂y 2 ∂y 4
 
∂ 4 ϕ1 4 ∂ 4 ϕ1 3 ∂ 4 ϕ1 2 2 ∂ 4 ϕ1 3 ∂ 4 ϕ1 4
= m +4 3 m m2 + 6 2 2 m1 m2 + 4 m1 m2 + m ξ3
∂ξ14 1 ∂ξ1 ∂ξ2 1 ∂ξ1 ∂ξ2 ∂ξ1 ∂ξ23 ∂ξ24 2

216
 
∂ 3 ϕ1 3 ∂ 3 ϕ1 2 ∂ 3 ϕ1 2 ∂ 3 ϕ1 3
+4 m +3 2 m m2 + 3 m1 m2 + m m3
∂ξ13 1 ∂ξ1 ∂ξ2 1 ∂ξ1 ∂ξ22 ∂ξ23 2

∂ 4 ϕ1 2 2 ∂ 4 ϕ1 2
+2 4 m n
1 1 + 3 2m1 n1 n2 + 2m1 m2 n21
∂ξ1 ∂ξ1 ∂ξ2

∂ 4 ϕ1 2 2
+ 2 2 m1 n2 + m22 n21 + 4m1 m2 n1 n2
∂ξ1 ∂ξ2

∂ 4 ϕ1 ∂ 4ϕ
1
+ (2m21 n1 n2 + 2m1 m2 n22 ) + m2 n 2 ξ 3
∂ξ1 ∂ξ23 ∂ξ24 2 2

∂ 3 ϕ1
+ (2m21 n1 n3 + 2m1 n21 m3 )
∂ξ13
∂ 3 ϕ1 2
+ 2 2m1 n2 n3 + 2n21 m2 m3 + 4m1 m2 n1 n3 + 4m1 m3 n1 n2
∂ξ1 ∂ξ2
∂ϕ1 2
+ 2 2m2 n1 n3 + 4m1 m2 n2 n3 + 4m2 m3 n1 n1 + 2m1 n22 m3
∂ξ1 ξ2

∂ 3 ϕ1 2 2

+ 3 2m2 n2 m3 + 2m2 n2 n3
∂ξ2

∂ 4 ϕ1 4 ∂ 4 ϕ1 3 ∂ 4 ϕ1 2 2
+ n + 4 n n 2 + 6 n n
∂ξ14 1 ∂ξ13 ∂ξ2 1 ∂ξ12 ∂ξ22 1 2

∂ 4 ϕ1 ∂ 4ϕ
1
+4 n1 n32 + n 4 ξ3
∂ξ1 ∂ξ23 ∂ξ24 2
 
∂ 3 ϕ1 3 ∂ 3 ϕ1 2 ∂ 3 ϕ1 2 ∂ 3 ϕ1 3
+4 n +3 2 n n2 + 3 n1 n2 + n n3 .
∂ξ13 1 ∂ξ1 ∂ξ2 1 ∂ξ1 ∂ξ22 ∂ξ23 2
Whence we have the following sufficient conditions for equality (4.7) to hold:
⎧ 4

⎪ m + 2m21 n21 + n41 = 0,
⎪ 1


⎪ 4m31 m2 + 4 m21 n1 n2 + m1 m2 n21 + 4n31 n2 = 0,



⎪ 2 2

⎪ 6m 2 2
m + 2 m n + m 2 2
n + 8m1 m2 n1 n2 + 6n21 n22 = 0,

⎪ 1 2

1 2 2 1




⎪ 4m1 m32 + 4 m22 n1 n2 + m1 m2 n22 + 4n1 n32 = 0,



⎪ 4 2 2 4

⎨ m2 + 2m2 n2 + n2 = 0,
4m31 m3 + 4 m21 n1 n3 + m1 n21 m3 + 4n31 n3 = 0,



⎪ 12m21 m2 m3 + 4 m21 n2 n3 + n21 m2 m3 +





⎪ + 8 (m1 m2 n1 n3 + m1 m3 n1 n2 ) + 12n21 n2 n3 = 0,





⎪ 12m1 m22 m3 + 4 m22 n1 n3 + m1 n22 m3 +



⎪ + 8 (m1 m2 n2 n3 + m2 m3 n1 n2 ) + 12n1 n22 n3 = 0,




4m32 m3 + 4 m2 m3 n22 + m22 n2 n3 + 4n32 n3 = 0,

217
which are equivalent to the following system:
⎧ 2

⎪ m21 + n21 = 0,

⎪ 2

⎪ m1 + n21 (m1 m2 + n1 n2 ) = 0,



⎪ 2

⎪ m1 + n21 m22 + n22 + 2 (m1 m2 + n1 n2 )2 = 0,


⎪ m2 + n2 (m m + n n ) = 0,



⎨ 2 2 1 2 1 2
2 2
⎪ m2 + n22 = 0,



⎪ m21 + n21 (m1 m3 + n1 n3 ) = 0,

⎪ 2

⎪ 2

⎪ m + n (m2 m3 + n2 n3 ) + 2 (m1 m2 + n1 n2 ) (m1 m3 + n1 n3 ) = 0,

⎪ 2
1 1


⎪ 2

⎪ m2 + n2 (m1 m3 + n1 n3 ) + 2 (m1 m2 + n1 n2 ) (m2 m3 + n2 n3 ) = 0,

⎩ 2
m2 + n22 (m2 m3 + n2 n3 ) = 0.

This system has the solution

n1 = ±m1 , n2 = ±im2 , n3 = ∓im3 , ∀ m1 , m2 , m3 ∈ C

or
n1 = ±m1 , n2 = 0, n3 = ∓im3 , m2 = 0, ∀ m1 , m3 ∈ C.
Therefore,
ξ1 = xm1 ± iym1 , ξ2 = xm2 ± iym2 , ξ3 = xm3 ∓ iym3 ,
where the signs ± are simultaneously either + or −, and

ϕ̃(x, y) = ϕ1 (x ± iy) · (x ∓ iy).

In such a way, we can obtain sufficient conditions for the coefficients ms , ns , s = 1, 2, 3, for the
addition ϕ2 (ξ1 , ξ2 ) to satisfy equation (4.7). For this purpose, we also calculate all necessary partial
derivatives:
∂ 4 ϕ2 4
4 ∂ ϕ2 ∂ 4 ϕ2 3
= m 1 + 4 m m2
∂x4 ∂ξ14 ∂ξ13 ∂ξ2 1
∂ 4 ϕ2 2 2 ∂ 4 ϕ2 3 ∂ 4 ϕ2 4
+6 m m + 4 m 1 m + m ,
∂ξ12 ∂ξ22 1 2 ∂ξ1 ∂ξ23 2
∂ξ24 2

∂ 4 ϕ2 ∂ 4 ϕ2 2 2 ∂ 4 ϕ2 2 2

= 4 m 1 n 1 + 3 2m 1 n 1 n 2 + 2m 1 m 2 n 1
∂x2 ∂y 2 ∂ξ1 ∂ξ1 ∂ξ2
∂ 4 ϕ2 2 2
+ 2 2 m1 n2 + 4m1 m2 n1 n2 + m22 n21
∂ξ1 ∂ξ2
∂ 4 ϕ2 2 2
∂ 4 ϕ2 2 2
+ 2m1 m 2 n 2 + 2m2 n1 n2 + m n ,
∂ξ1 ∂ξ23 ∂ξ24 2 2

∂ 4 ϕ2 4
4 ∂ ϕ2 ∂ 4 ϕ2 3
= n 1 + 4 n n2
∂y 4 4
∂ξ1 ∂ξ13 ∂ξ2 1
∂ 4 ϕ2 2 2 ∂ 4 ϕ2 3 ∂ 4 ϕ2 4
+6 n n + 4 n n
1 2 + n .
∂ξ12 ∂ξ22 1 2 ∂ξ1 ∂ξ23 ∂ξ24 2

218
Substituting the obtained quantities into the biharmonic equation, we obtain the system
⎧ 4

⎪ m1 + 2m21 n21 + n41 = 0,



⎪ 3 2 2 3

⎨ 4m1 + 4m1 n1 n2 + 4m1 m2 n1 + 4n1 n + 2 = 0,
6m21 m22 + 2m21 n22 + 8m1 m2 n1 n2 + 2m22 n21 + 6n21 n22 = 0,



⎪ 4m1 m32 + 4m1 m2 n22 + 4m22 n1 n2 + 4n1 n32 = 0,



⎩ m4 + 2m2 n2 + n4 = 0,
2 2 2 2

which is equivalent to the following systems


⎧ 2 2

⎪ m1 + n21 = 0,

⎪ 2

⎪ 2

⎨ m 1 + n 1 (m1 m2 + n1 n2 ) = 0,

m21 + n21 m22 + n22 + 2 (m1 m2 + n1 n2 )2 = 0,



⎪ 2 2

⎪ m 1 + n 2 (m1 m2 + n1 n2 ) = 0,

⎩ 2
m22 + n22 = 0

and ⎧ 2 2

⎨ m1 + n1 = 0,
m22 + n22 = 0,


m1 m2 + n1 n2 = 0,
whose solutions are of the form
m1 = ±in1 , m2 = ±in2 .
If the signs ± are chosen to be different, then we have trivial cases when n1 = 0 or n2 = 0 and ξ2 = 0.
It is obvious that we obtain the holomorphic function ϕ2 (ξ1 , ξ2 ) ≡ ϕ2 (ξ1 ) = ϕ2 (xm1 ± iym1 ) that
satisfies the biharmonic equation. If the signs ± are chosen to be identical, then we have the identity
m1 m2 + n1 n2 = 0 whose solution is

n1 = im1 , n2 = im2 , ∀ m1 , m2 ∈ C.

In this case,
ξ1 = xm1 + iym1 , ξ2 = xm2 + iym2
and the function
ϕ2 (ξ1 , ξ2 ) = ϕ2 (xm1 + iym2 , xm2 + iym2 ) = ϕ2 (x + iy)
is a holomorphic function of a complex variable, which also satisfies the biharmonic equation.
Thus, after all transformations, we obtain the function

f3 (x, y) = ϕ1 (x ± iy) · (x ∓ iy) + ϕ2 (x ± iy), (4.8)

where ϕ1 and ϕ2 are holomorphic functions of a complex variable.


Obviously, formula (4.8) gives a general solution of equation (4.7), and it is the well-known Gour-
sat formula. Therefore, every H-analytic function is a solution of the Laplace equation (4.1) and
biharmonic equation (4.7) for the corresponding choice of the coefficients ξ1 , ξ2 , ξ3 .

219
REFERENCES
1. F. Hausdorff, “Zur Theorie der Systeme complexer Zahlen,” Leipziger Berichte, 52, 43–61 (1900).
2. F. Ringleb, “Beiträge zur funktionentheorie in hyperkomplexen systemen, I.” Rend. Circ. Mat. Palermo,
57(1), 311–340 (1933).
3. R. F. Rinehart and J. C. Wilson, “Two types of differentiability of functions on algebras,” Rend. Circ.
Matem. Palermo, II(11), 204–216 (1962).
4. W. O. Portman, “A derivative for Hausdorff-analytic functions,” Proc. Amer. Math. Soc., V(10), 101–105
(1959).
5. M. E. Luna-Elizarraras, M. Shapiro, and V. Shpakivskyi, “On the Hausdorff Analyticity for Quaternion-
Valued Functions,” Complex Anal. Oper. Theory, 13, 2863–2880 (2019).
6. S. N. Volovel’skaya, “The experience of construction of elements of the theory of functions in a commutative
associative system with three units,” Zapiski Nauchno-Issledovatel’skogo Instituta Matematiki i Mehaniki i
Har’kovskogo Matematicheskogo Obshchestva, 16, 143–157 (1939).
7. D. Alpay, M. E. Luna-Elizarraras, M. Shapiro, and D. C. Struppa, Basics of functional analysis with
bicomplex scalars, and bicomplex schur analysis. SpringerBriefs in Mathematics, Springer, 2014.
8. M. E. Luna-Elizarraras, M. Shapiro, D. C. Struppa, and A. Vajiac, Bicomplex holomorphic functions: the
algebra, geometry and analysis of bicomplex numbers, Springer, 2015.
9. T. Kuzmenko and V. Shpakivskyi, “G-monogenic mappings in a three-dimensional noncommutative alge-
bra,” Complex Variables and Elliptic Equations. DOI: 10.1080/17476933.2021.1947257 (2021).
10. T. Kuzmenko and V. Shpakivskyi, “Differentiable functions in a three-dimensional associative noncommu-
tative algebra,” Advances in the Theory of Nonlinear Analysis and its Application, 1(6), 66–73 (2022).
11. G. Rota, B. Sagan, and P. Stein, “A cyclic derivative in noncommutative algebra,” Journal of Algebra, 64,
54–75 (1980).
12. S. V. Gryshchuk and S. A. Plaksa, “Monogenic functions in the biharmonic boundary value problem,”
Math. Meth. Appl. Sci., 39(11), 2939–2952 (2016).
13. S. V. Gryshchuk and S. A. Plaksa, “Reduction of a Schwartz-type boundary value problem for biharmonic
monogenic functions to Fredholm integral equations,” Open Mathematics, 15(1), 374–381 (2017).
14. I. P. Mel’nichenko and S. A. Plaksa, Commutative algebras and spatial potential fields (in Russian). Kiev,
Inst. Math. NAS Ukraine, 2008.
15. V. S. Shpakivskyi, “Hypercomplex method for solving linear PDEs,” Proc. of the IAMM NAS of Ukraine,
32, 160–181 (2018).
16. V. S. Vladimirov, Methods of the theory of functions of several complex variables [in Russian]. Moscow,
Nauka, 1964.
17. V. S. Shpakivskyi, “Constructive description of monogenic functions in a finite-dimensional commutative
associative algebra,” Adv. Pure Appl. Math., 7(1), 63–76 (2016).
18. S. A. Plaksa and R. P. Pukhtaievych, “Monogenic functions in a finite-dimensional semi-simple commutative
algebra,” An. Şt. Univ. Ovidius Constanţa, 22(1), 221–235 (2014).

220
19. A. Pogorui, R. M. Rodriguez-Dagnino, and M. Shapiro, “Solutions for PDEs with constant coefficients
and derivability of functions ranged in commutative algebras,” Math. Meth. Appl. Sci., 37(17), 2799–2810
(2014).

Vitalii Shpakivskyi
Institute of Mathematics of the NAS of Ukraine, Kyiv, Ukraine
E-Mail: shpakivskyi86@gmail.com
Tetiana Kuzmenko
Zhytomyr Military Institute, Zhytomyr, Ukraine
E-Mail: kuzmenko.ts15@gmail.com

221

You might also like