Inverse problems of recovering lower-order coefficients from

boundary integral data

1
S.G. Pyatkov, O.A. Soldatov
arXiv:2412.15635v1 [math.AP] 20 Dec 2024

Abstract. Under consideration are mathematical models of

heat and mass transfer. We study inverse problems of re-
covering lower-order coefficients in a second order parabolic
equation. The coefficients are representable in the form of a
finite segments of the series whose coefficients depending on
time are to be determined. The linear case is also considered.
The overdetermination conditions are the integrals over the
boundary of the domain of a solution with weights. The
main attention is paid to existence, uniqueness, and stability
estimates for solutions to inverse problems of this type. The
problem is reduced to an operator equation which is studied
with the use of the fixed point theorem and a priori esti-
mates. A solution has all generalized derivatives occurring
into the equation summable to some power. The method of
the proof is constructive and it can be used for developing
new numerical algorithms of solving the problem.

1 Introduction
We examine the question on identification of lower-order coefficients and the
right-hand side in a parabolic equation. An equation is of the form

Mu = ut + Au = f (t, x), (t, x) ∈ Q = (0, T ) × G, (1.1)

where G ∈ Rn is a bounded domain with boundary Γ. Let S = (0, T ) × Γ.

The function f and a second order elliptic operator A are representable as
r
X s
X
A(t, x, D) = A0 (t, x, Dx )+ qi (t)Ai (t, x, Dx ), f = f0 (t, x)+ fi (t, x)qi (t),
i=1 i=r+1
n
X n
X n
X
A0 = − akl (t, x)∂xk xl + ak (t, x)∂xk +a0 , Ai = aik (t, x)∂xk +ai0 , i = 1, 2, . . . , r.
k,l=1 k=1 k=1

1
MSC Primary 35R30; Secondary 35K20; 80A20
Keywords: inverse problem; heat transfer coefficient; convection-diffusion equation; heat
and mass transfer; integral measurements

1
The equation (1.1) is furnished with the initial and boundary conditions
n
∂u ∂u X
u|t=0 = u0 , Bu|S = + σu = g(t, x), = γi (t, x)uxi . (1.2)
∂γ ∂γ i=1

The additional conditions to determine the coefficients are as follows:

Z
u(t, x)ϕj (x) = ψj (t), j = 1, 2, . . . , s. (1.3)
Γ

The unknowns in the problem (1.1)-(1.3) are a solution u and the functions
qi (t) (i = 1, 2, . . . , s). The problems (1.1)-(1.3) arise when describing heat
and mass transfer processes, diffusion and filtration processes, in ecology and
many other fields [1], [2]. The overdetermination conditions (1.3) are not
standard. We can refer to the monograph [3] devoted to inverse parabolic
problems and the monographs [4, 5, 6] containing basic statements of inverse
problems including those for parabolic equations. In the monograph [4] (see
also the results in [7]), inverse problems of recovering coefficients independent
of some spatial variables are studied and existence and uniqueness theorems
are justified. In view of the method, the known coefficients also are indepen-
dent of some spatial variables. The overdetermination conditions are values
of a solution on some hyperplanes. More complete results are exposed in [8]-
[10], where well-posedness of inverse problems of recovering coefficients from
the values of a solutions on some manifolds are studied. Inverse problems
with pointwise or integral data of the form (1.3) (in the latter case only the
problem of identification of the heat transfer coefficient σ is considered in
[11]) are studied in the articles by Prilepko A.I. and a series of interesting
problems is described in [3]. Similar results but with other conditions on the
data and in other spaces are obtained in [12]-[15]. Integral overdetermina-
tion conditions with integrals over the domain G are used in [16]-[18], where
either the heat transfer coefficient or the flux on the boundary are to be de-
termined. These conditions is also often involved in (1.3) [11], [19, 20], where
some numerical methods for solving the problem are also exposed. In [21]
the problem of simultaneous recovering the heat transfer coefficient which
depends on an integral of a solution with weight and lower-order coefficient
with the use of the condition (1.3) and the integral of a solution over the
domain with weight are studied. Integral overdetermination conditions (1.3)
are often used as some approximations of pointwise conditions. It is noted,
in particular, in [19, 20]. Numerical solution of the problems (1.1)-(1.3) are

2
also discussed in [22]-[23]). The problem of recovering the heat transfer coef-
ficient with the overdetermination condition (1.3) is considered in [24]. Note
that we do not know articles devoted to the problem (1.1)-(1.3) except for
our article [25]. The main attention is paid to the case of recovering coeffi-
cients of higher order derivatives which are sought in the class of continuous
functions. The existence and uniqueness theorems are obtained. But if we
deal with real problems of heat and mass transfer then it is more interesting
to consider the case of nonsmooth coefficients. By the same approach, we
obtain here a similar result for the lower-order coefficients which belong to
some Lebesgue space. Note that the integral equations to which these two
problems are reduced do not coincide. This

2 Preliminaries
Let E be a Banach space. By Lp (G; E) (G is a domain in Rn ), we mean the
space of measurable functions defined on G with values in E and the finite
norm kku(x)kE kLp (G) [26]. We use the Sobolev spaces Wps (G; E), Wps (Q; E)
(see [27, 28]) and the Hölder spaces C α,β (Q), C α,β (S) (see [29]). By a norm
of a vector, we mean the sum of the norms of coordinates. Given an interval
J = (0, T ), put Wps,r (Q) = Wps (J; Lp (G)) ∩ Lp (J; Wpr (G)). Respectively,
Wps,r (S) = Wps (J; Lp (Γ)) ∩ Lp (J; Wpr (Γ)). All function spaces and coefficients
of the equation (1.1) are assumed to be real. Next, we suppose that p > n+2
and Γ ∈ C 3 (see the definition in [29]). Given sets X, Y , the symbol ρ(X, Y )
stands for the distance between them. Introduce the notations Qτ = (0, τ ) ×
G, S τ = (0, τ ) × Γ. Gδ = {x ∈ G : ρ(x, Γ) < δ}, Qδ = (0, T ) × Gδ ,
Qτδ = (0, τ ) × Gδ . Construct a function ϕ(x) ∈ C ∞ (G) such that ϕ(x) = 1 in
Gδ/2 and ϕ(x) = 0 in G\G3δ/4 . A small parameter δ > 0 (it can be arbitrarily
small) is fixed below. It characterizes a neighborhood of the boundary, where
additional smoothness requirements are imposed on the data. Consider the
auxiliary problem
n
X n
X
Mu = ut + A0 u = f0 (t, x), A0 u = − aij uxi ,xj + ai uxi + a0 u, (2.1)
i,j=1 i=1

∂u
+ σu = g, u|t=0 = u0 , (2.2)
∂γ S

3
We assume that
u0 ∈ Wp2−2/p (G), B(0, x, D)u0|Γ = g(0, x), g ∈ Wps0 ,2s0 (S), f0 ∈ Lp (Q), (2.3)
where s0 = 1/2 − 1/2p;
ϕu0 (x) ∈ Wp3−2/p (G), ϕf0 ∈ Lp (0, T ; Wp1(G), g ∈ Wps1 ,2s1 (S), (2.4)
where s1 = 1 − 1/2p;
aij ∈ C(Q), γi , σ ∈ Wps0 ,2s0 (S), ak ∈ Lp (Q), i, j = 1, 2, . . . , n, k ≤ n; (2.5)
1
aij ∈ L∞ (0, T ; W∞ (Gδ )), ak ∈ L∞ (0, T ; Wp1(Gδ )), γi , σ ∈ Wps1 ,2s1 (S); (2.6)
where i, j = 1, . . . , n k=0,1,. . . ,n, l=1,2,. . . ,r. Next, we suppose that
n
X
| γi νi | ≥ ε0 > 0, ∀(t, x) ∈ S, (2.7)
i=1

where ν is the outward unit normal to Γ and ε0 is a positive constant. The

operator A0 is elliptic, i. e., there exists a constant δ0 > 0 such that
n
X
aij ξi ξj ≥ δ0 |ξ|2 ∀(t, x) ∈ Q, ∀ξ ∈ Rn . (2.8)
i,j=1

The next theorem is a consequence of Theorem 1 in [30].

Theorem 2.1 Let the conditions (2.3), (2.5), (2.7), (2.8) hold. Then there
exists a unique solution u ∈ Wp1,2 (Q) to the problem (2.1), (2.2). The follow-
ing estimates is valid:
 
kukWp1,2(Q) ≤ c ku0 kWp2−2/p (G) + kf0 kLp (Q) + kgkWps0,2s0 (S) . (2.9)

If additionally the conditions (2.4), (2.6) hold then a solution is such that
ϕut ∈ Lp (0, T ; Wp1(G)), ϕu ∈ Lp (0, T ; Wp3(G)). In the case of g = 0 and
u0 = 0 we have the estimates
kukWp1,2 (Qτ ) ≤ ckf kLp (Qτ ) , (2.10)

kukWp1,2(Qτ ) + kϕut kLp (0,τ ;Wp1 (G)) + kϕukLp (0,τ ;Wp3 (G)) ≤
c(kf0 kLp (Qτ ) + kϕf0 kLp (0,τ ;Wp1 (G)) ), (2.11)
where the constant c is independent of f, τ ∈ (0, T ].

4
3 Existence and uniqueness theorems
Proceed with additional conditions on the data. We assume that
1 1
Z
1
ψi ∈ Wp (0, T ), ψi (0) = u0 ϕi dΓ, ϕi ∈ Lq (Γ), + = 1, (3.1)
Γ q p
where i = 1, . . . , s,

fm ∈ L∞ (0, T ; Lp (G)) ∩ L∞ (0, T ; Wp1(Gδ )) ∩ C([0, T ]; Lp (Γ)), (3.2)

where m = r + 1, . . . , s,

alk ∈ L∞ (0, T ; Wp1(Gδ )) ∩ L∞ (0, T ; Lp (G)) ∩ C([0, T ]; Lp (Γ)), (3.3)

for all l = 1, . . . , r, k = 0, 1, . . . , n. Assuming that the conditions of The-

orem 1 and the above conditions are fulfilled, we can construct a solu-
tion to the problem (2.1), (2.2). It possesses the properties Φ ∈ Wp1,2 (Q),
ϕΦt ∈ Lp (0, T ; Wp1(G)), ϕΦ ∈ Lp (0, T ; Wp3(G)). Consider the matrix B0 of
dimension s × s with the rows

− < A1 (t, x, D)Φ, ϕj >, . . . , − < Ar (t, x, D)Φ, ϕj >,

< fr+1 (t, x), ϕj >, . . . , < fs (t, x), ϕj >, j ≤ s,
R
where < u, v >= Γ u(x)v(x) dΓ. The embedding theorems imply that Φ ∈
C 1−(n+2)/2p,2−(n+2)/p (Q) (see Lemma 3.3, Ch.2 [29]). In this case the entries
of the matrix B0 are continuous functions. We require that

det B0 6= 0 ∀t ∈ [0, T ]. (3.4)

Since in Theorem 3.1 below we prove existence of solutions locally in time,

the condition (3.4) can be replaced with the condition

det B 6= 0, ∀t ∈ [0, T ]. (3.5)

where the matrix B has the rows

− < A1 (t, x, D)u0 , ϕj >, . . . , − < Ar (t, x, D)u0, ϕj >,

< fr+1 (t, x), ϕj >, . . . , < fs (t, x), ϕj >, j ≤ s.

5
This condition is easier to check in contrast to that in (3.4). If we replace in
Theorem 3.1 the condition (3.4) with the condition (3.5) then the claim of
Theorem 3.1 remains valid.
Make the change of variables u = v + Φ. The problem (1.1)-(1.3) is
reduced to an equivalent problem
s
X r
X
vt + A(~q)v = qi fi − qi Ai Φ = f 1 , (3.6)
i=r+1 i=1
Pr
where A(~q)v = A0 v + A1 (~q)v, A1 (~q) = i=1 qi Ai ;

v|t=0 = 0, Bv|S = 0; (3.7)

< v, ϕj >= ψj (t)− < Φ, ϕj >= ψ̃j , i = 1, . . . , s. (3.8)

Theorem 3.1 Let the conditions (2.3)-(2.8), (3.1)-(3.4) hold. Then there
exists a number τ0 ∈ (0, T ] such that on the segment [0, τ0 ] there exists a
unique solution (u, q1 , q2 , . . . , qs ) to the problem (1.1)-(1.3) such that u ∈
Wp1,2 (Qτ0 ), ϕu ∈ Lp (0, τ0 ; Wp3 (G)), ϕut ∈ Lp (0, τ0 ; Wp1 (G)), qj ∈ Lp (0, τ0 ),
j = 1, 2, . . . , s.

Proof. We have reduced our problem (1.1)-(1.3) to an equivalent simpler

problem (3.6)- (3.8). Denote by Hτ the space of functions v satisfying the
conditions (3.7) and such that v ∈ Wp1,2 (Qτ ), ϕvt ∈ Lp (0, τ ; Wp1 (G)), ϕv ∈
Lp (0, τ ; Wp3 (G)). Let kvkHτ = kvkWp1,2(Qτ ) +kϕvt kLp (0,T ;Wp1 (G)) +kϕvkLp (0,T ;Wp3(G)) .
Define also the space Wτ of functions f ∈ Lp (Qτ ) such that ϕf ∈ Lp (0, τ ; Wp1 (G)).
Endow this space with the norm kf kWτ = kf kLp (Qτ ) + kϕf kLp (0,τ ;Wp1 (G)) . In
view of Theorem 1, for every function f 1 ∈ Wτ , there exists a unique solution
v = (∂t − A0 )−1 f 1 to the equation vt − A0 v = f 1 satisfying the conditions
(3.7) and the estimate
kvkHτ ≤ c1 kf 1 kWτ , (3.9)
where the constant c1 is independent of τ . Reduce our problem to an integral
equation. Multiply the equation (3.6) by ϕj and integrate over Γ. Note
that in the class of solutions u described in Theorem 2 traces of summands
occurring in (3.6) on Γ exist. We obtain the equality
r
X s
X
ψ̃j′ + < A(~q)v, ϕj >= − qi < Ai Φ, ϕj > + qi < fi , ϕj > . (3.10)
i=1 i=r+1

6
The right-hand side of this equality coincides with the j-th coordinate of the
vector B0 (t)~q. In this case the system (3.10) can be written in the form

~q(t) = B0−1 H(~q)(t) = R(~q), H(~q) = (ψ̃1′ + < A(~q)v, ϕ1 >,

ψ̃2′ + < A(~q)v, ϕ2 >, . . . , ψ̃s′ + < A(~q)v, ϕs >)T , (3.11)

where v is a solution to the problem (3.6), (3.7). Let ~q = 0. In this case the
right-hand side in (3.11) is written as follows:

~ Ψ
R(0) = B0−1 Ψ, ~ = (ψ̃ ′ , ψ̃ ′ , . . . , ψ̃ ′ )T .
1 2 s

Assign R0 = 2kR(0)kLp (0,T ) . Introduce that ball Bτ = {~q ∈ Lp (0, τ ) :

k~qkLp (0,τ ) ≤ R0 }. Next our aim is to prove that there exists a solution
to the equation (3.11). We employ the fixed point theorem. To this end, we
first obtain estimates for solutions to the problem (3.6), (3.7) assuming that
~q ∈ Bτ . We have that

v = −(∂t + A0 )−1 A1 (~q)v + (∂t + A0 )−1 f 1 . (3.12)

From (3.9), it follows that

k(∂t + A0 )−1 A1 (~q)vkHτ ≤ c2 kA1 (~q)vkWτ . (3.13)

The following estimate is valid in accord with our condition on the coeffi-
cients:
r
X
1
kA (~q)vkWτ ≤ k qj Aj (t, x, D)vkWτ
j=1
r
X
≤ k~qkLp (0,τ ) (kAj (t, x, D)vkL∞ (0,τ ;Lp (G)) + kϕAj (t, x, D)vkL∞ (0,τ ;Wp1 (G)) )
j=1

≤ c3 k~qkLp (0,τ ) (kvkL∞ (0,τ ;W∞

1 (G)) + kϕvkL (0,τ ;W 2 (G)) ).
∞ ∞
(3.14)

The constant c3 depends on the norm of coefficients in L∞ (0, τ ; Lp (G)) ∩

2−2/p
L∞ (0, τ ; Wp1 (Gδ )). Note that Wp1,2 (Qτ ) ⊂ C([0, τ ]; Wp (G)). Moreover, if
1 3 3−2/p
ϕvt ∈ Lp (0, τ ; Wp (G)), ϕv ∈ Lp (0, τ ; Wp (G)) then ϕu ∈ C([0, τ ]; Wp (G)).
These embeddings result from [28, theorem III 4.10.2]. Even more, in both

7
cases the embedding constant is indepedent of τ . Next, we have the estimate
(see Theorem 4.6.1, 4.6.2 [26])

kvkL∞ (0,τ ;W∞

1 (G)) + kϕvkL (0,τ ;W 2 (G)) ≤ c4 (kvk
∞ ∞ L∞ (0,τ ;Wp1+s (G)) +
θ
kϕvkL∞ (0,τ ;W∞
2+s
(G)) ) ≤ c5 (kvkL 2−2/p kvk1−θ
L∞ (0,τ ;Lp (G)) +
∞ (0,τ ;Wp (G))

kϕvkθ1 3−2/p kvkL1−θ 1

∞ (0,τ ;Lp (G))
), (3.15)
L∞ (0,τ ;Wp (G))

where n/p < s < 1 − 2/p, 1 + s = θ(2 − 2/p), 2 + s = θ1 (3 − 2/p) and we have
used the interpolation inequalities [26]. The Newton-Leibnitz formula yields
kvkL∞ (0,τ ;Lp (G)) ≤ τ (p−1)/p kvt kLp (Qτ ) . (3.16)
The estimates (3.14), (3.15), (3.16) ensure the inequality
kA1 (~q)vkWτ ≤ c6 k~qkLp (0,τ ) kvkHτ τ β , β = min(β0 , β1 ), (3.17)
where β0 = (1 − θ)(p − 1)/p, β1 = (1 − θ1 )(p − 1)/p. Choose τ0 from the
condition c6 R0 τ β = 1/2. In this case the relation (3.12) yields
kvkHτ ≤ 2k(∂t + A0 )−1 f 1 kHτ . (3.18)
In view of (3.9), we can write out the estimate
kvkHτ ≤ c7 kf 1 kWτ ≤ c8 k~qkLp (0,τ ) , τ ≤ τ0 , (3.19)
where we can assume that the constant c8 is independent of τ ≤ τ0 . It
depends on the norms of the data. Next, we consider two vectors q~i =
(q1i , q2i , . . . , qsi ), i = 1, 2, and construct the corresponding solutions vi to the
problem (3.6), (3.7). Subtracting two equations (3.6) we conclude that the
difference w = v1 − v2 is a solution to the problem

wt + A((~q1 + ~q2 )/2)w + A1 (~q1 − ~q2 )(v1 + v2 )/2 =

X s r
X
(q1i − q2i )fi − (q1i − q2i )Ai Φ, (3.20)
i=r+1 i=1

w|t=0 = 0, Bw|S = 0. (3.21)

As before, in view of the inequality (3.9), we have the estimate

kwkHτ ≤ c1 kA1 ((~q1 + ~q2 )/2)wkWτ +

c1 kA1 (~q1 − ~q2 )(v1 + v2 )/2kWτ + c10 k~q1 − ~q2 kLp (0,τ ) . (3.22)

8
Applying the estimate (3.17), we infer

kwkHτ ≤ c6 R0 τ β kwkHτ + c6 τ β k~q1 − ~q2 kLp (0,τ ) k(v1 + v2 )/2kHτ +

c10 k~q1 − ~q2 kLp (0,τ ) . (3.23)

Since τ ≤ τ0 , the last estimate implies that

kwkHτ ≤ 2c6 τ β k~q1 − ~q2 kLp (0,τ ) k(v1 + v2 )/2kHτ + 2c10 k~q1 − q~2 kLp (0,τ ) . (3.24)

Next, involving the estimate (3.19) written for the functions vi , we conclude
that
kv1 − v2 kHτ ≤ c11 k~q1 − ~q2 kLp (0,τ ) , (3.25)
with a constant c11 independent of τ ≤ τ0 . Estimate the norm kR(~q1 ) −
R(~q2 )kLp (0,τ ) with τ ≤ τ0 . Actually, the required estimate has been already
established. Consider the expressions Ij =< (A(~q1 )v1 − A(~q2 ))v2 , ϕj > oc-
curring in the difference R(~q1 ) − R(~q2 ). The Hölder inequality yields

|Ij | = | < A0 w + A1 ((~q1 + q~2 )/2)w + A1 (~q1 − ~q2 )(v1 + v2 )/2, ϕj > | ≤
c12 (kA0 wkLp (Γ) + kA1 ((~q1 + ~q2 )/2)wkLp (Γ) + kA1 (~q1 − ~q2 )(v1 + v2 )/2kLp (Γ) .

The second summand on the right-hand side of Ij is estimated as follows (see

(3.17)):

kA1 ((~q1 + ~q2 )/2)wkLp(0,τ ;Lp (Γ)) ≤ c12 kA1 ((~q1 + ~q2 )/2)wkWτ
≤ c13 τ β R0 kwkHτ . (3.26)

In view of the inequality (3.19) written for the functions vi , the third sum-
mand admits the estimate

(v1 + v2 ) (v1 + v2 )
kA1 (~q1 − ~q2 ) kLp (0,τ ;Lp (Γ)) ≤ c14 kA1 (~q1 − ~q2 ) kW τ ≤
2 2
c13 τ β C1 (R0 )k~q1 − ~q2 kLp (0,τ ) . (3.27)

At last, the first summand is estimated as follows:

kA0 wkLp (0,τ ;Lp (Γ)) ≤ c15 kϕA0 wkLp (0,τ ;Wps (G)) ≤
c16 kϕA0 wksLp (0,τ ;Wp1 (G)) kϕA0 wk1−s
Lp (Qτ ) , (3.28)

9
where n/p < s < 1. Here we have used the interpolation inequality [26].
Using the conditions on the coefficients, we obtain the estimate

kϕA0 wkLp (G) ≤ c16 (kϕwkWp2(G) + kwkL∞ (0,τ ;W∞

1 (G) ). (3.29)

Next, we repeat the arguments used in proof of the estimate (3.17) (see the
inequality (3.15)). Finally, we arrive at the estimate

kϕA0 wkLp (G) ≤ c17 τ β2 kϕwkHτ , β2 > 0. (3.30)

The estimates (3.26), (3.27), (3.30), (3.25) ensure the estimate

kR(~q1 ) − R(~q2 )kLp (0,τ ) ≤ c18 τ β3 k~q1 − ~q2 kLp (0,τ ) , (3.31)

where β3 is a positive constant and c18 is a constant depending on R0 but

independent of τ . Choose a quantity τ1 ≤ τ0 such that c18 τ β3 ≤ 1/2 for
τ ≤ τ1 . In this case the operator R(~q) takes the ball Bτ1 into itself and
is a contraction. The fixed point theorem implies that the equation (3.11)
has a unique solution in the ball Bτ1 . Let v = v(~q). Demonstrate that this
functions satisfies (3.8). Integrating (3.6) with the weight ϕj over Γ, we
obtain the equalities
r
X
< vt , ϕj > + < A(~q)v, ϕj >= − qj < Aj Φ, ϕj >
j=1
s
X
+ < fj , ϕj > qj (t). (3.32)
j=r+1

Subtracting them from (3.10), we conclude that < vt , ϕj > −ψ̃j′ = 0 for all j
and thereby the conditions (3.8) hold. Uniqueness of solutions follows from
the estimates obtained in the above proof.
Remark. Stability estimate for solutions also holds.
In the linear case, i,e, the unknown functions occur only into the right-
hand side the claim becomes global in time. In this case A = A0 and the
matrix B0 has the rows

< f1 (t, x), ϕj >, . . . , < fs (t, x), ϕj >, j ≤ s, (3.33)

i.e., r = 0.
More exactly, the following theorem is valid.

10
Theorem 3.2 Let the conditions (2.3)-(2.8), (3.1), (3.2), (3.4) hold. Then
there exists a unique solution (u, q1 , q2 , . . . , qs ) to the problem (1.1)-(1.3) such
that u ∈ Wp1,2 (Q), ϕu ∈ Lp (0, T ; Wp3(G)), ϕut ∈ Lp (0, T ; Wp1(G)), qj ∈
Lp (0, T ), j = 1, 2, . . . , s.

The proof is in lines with that of the main result in [10]. The idea of
the proof is as follows. First, we prove solvability of the problem on some
small segment of time [0, τ0 ]. Next, repeating the arguments we establish
solvability on [τ0 , τ1 ], and so on. Since the problem is linear, it is possible to
prove that the length τi −τi−1 does not tends to zero which implies solvability
on the whole segment [0, T ].

4 Discussion
Under consideration are mathematical models of heat and mass transfer. We
study inverse problems of recovering lower-order coefficients and the right-
hand side in a second order parabolic equation. The coefficients are repre-
sentable in the form of a finite segments of series whose coefficients depending
on time are to be determined. The linear case in which only the right-hand
side is recovered is also considered. The overdetermination conditions are
the integrals over the boundary of a domain of a solution with weights. The
main attention is paid to existence, uniqueness, and stability estimates for
solutions to inverse problems of this type. The problem is reduced to an
operator equation which is studied with the use of the fixed point theorem
and a priori estimates. A solution has all generalized derivatives occurring
into the equation summable to some power. The method of the proof is
constructive and it can be used for developing new numerical algorithms for
solving the problem.

5 Conclusions
The existence and uniqueness theorems in inverse problems of recovering the
lower-order coefficients in a parabolic equation from the boundary integral
measurements are proven locally in time. They are sought in the form of a
finite segment of the Fourier series with coefficients depending on time. The
proof relies on a priori bounds and the fixed point theorem. The conditions
on the data ensuring existence and uniqueness of solutions in Sobolev classes

11
are sharp. They are smoothness and consistency conditions on the data
and additional conditions on the kernels of the integral operators used in
additional measurements.
Acknowledgement. The research was carried out within the state assign-
ment of Ministry of Science and Higher Education of the Russian Federation
(theme No. FENG-2023-0004, ”Analytical and numerical study of inverse
problems on recovering parameters of atmosphere or water pollution sources
and (or) parameters of media”).

References
[1] Alifanov, O.M.; Artyukhin, E.A.; Nenarokomov, A.V. Inverse Prob-
lems in the Study of Complex Heat Transfer; Janus-K: Moscow, Russia,
2009.

[2] Ozisik, M.N.; Orlande, H.R.B. Inverse Heat Transfer; Taylor & Francis:
New York, USA 2000.

[3] Prilepko, A.I., Orlovsky, D.G., Vasin, I.A. Methods for Solving Inverse
Problems in Mathematical Physics; Marcel Dekker: New York, 1999.

[4] Belov, Ya.Ya. Inverse Problems for Parabolic Equations; VSP: Utrecht,
2002.

[5] Isakov, V. Inverse Problems for Partial Differential Equations;

Springer: Berlin, 2006.

[6] Kabanikhin, S.I. Inverse and Ill-Posed Problems; Walter de Gruyter:

Berlin, Boston, 2012.

[7] Frolenkov, I.V., Romanenko, G.V. On Solving an Inverse Problem for

a Multidimensional Parabolic Equation, Siberian Journal of Industrial
Mathematics, 2012, 15, no. 2(50), 139–146.

[8] Pyatkov, S.G., Samkov, M.L. On Some Classes of Coefficient Inverse

Problems for Parabolic Systems of Equations, Sib. Adv. in Math.,
2012, 22(4), 287–302.

12
 [9] Pyatkov, S.G., Tsybikov, B.N. On Some Classes of Inverse Problems
for Parabolic and Elliptic Equations J. Evol. Equat., 2011, 11(1), 155–
186.

[10] Pyatkov, S.G. On Some Classes of Inverse Problems for Parabolic

Equations, J. Inv. Ill-Posed problems, 2011, 18(8), 917–934.

[11] Kostin, A.B., Prilepko, A.I. On Some Problems of the Reconstruction

of a Boundary Condition for a Parabolic Equation, II, Differ. Equat.,
1996, 32, 1515–1525.

[12] Pyatkov, S.G. On Some Classes of Inverse Problems with Overdeter-

mination Data on Spatial Manifolds, Siberian Mathematical Journal,
2016, 57(5), 870–880.

[13] Pyatkov, S.G., Rotko, V.V. On some parabolic inverse problems with
the pointwise overdetermination Siber. Adv. in Math., 2020, 30(2),
124–142.

[14] Pyatkov, S.G. Identification of Thermophysical Parameters in Mathe-

matical Models of Heat and Mass Transfer, Journal of Computational
and Engineering Mathematics, 2022, 9(2), 52–66.

[15] Pyatkov, S.G., Rotko, V.V. Inverse Problems with Pointwise Overde-
termination for some Quasilinear Parabolic Systems, AIP Conference
Proceedings, 2017, 1907, 020008.

[16] Pyatkov, S., Soldatov, O., Fayazov, K. Inverse problems of recovering

the heat transfer coefficient with integral data, J. of Math. Sci., 2023,
274(2), 255–268.

[17] Kozhanov, A.I. Linear Inverse Problems for Some Classes of Nonlinear
Nonstationary Equations, Siberian Electronic Reports, 2015, 12, 264–
275.

[18] Verzhbitskiy, M. A., Pyatkov, S. G. On Some Inverse Problems of Re-

covering Boundary Regimes, Math. Notes of NEFU, 2016, 23(2), 3–18.

[19] Dihn, N.; Hao, D.N.; Thanh, P.X.; Lesnik, D. Determination of the
Heat Transfer Coefficients in Transient Heat Conduction. Inverse Prob-
lems, 2013, 29, 095020.

13
[20] Hao, D.N.; Huong, B.V.; Thanh, P.X.; Lesnik, D. Identification of
Nonlinear Heat Transfer Laws from Boundary Observations, Applicable
Analysis, 2014, 94(9), 1784–1799.

[21] Lorenzi, A.; Messina, F.; Identifying a Spherically Symmetric Conduc-

tivity in a Nonlinear Parabolic Equation, Applicable Analysis, 85(8),
867–889.

[22] Samarskii, A.A.; Vabishchevich P.N. Numerical Methods for Solving

Inverse Problems of Mathematical Physics. Walter de Gruyter GmbH
& Co. KG: Berlin/Boston, 2007.

[23] Huntul, M.J. Identification of the Timewise Thermal Conductivity in a

2D Heat Equation from Local Heat Flux Conditions, Inverse Problems
in Science and Engineering, 2021, it 29(7), 903–919.

[24] Pyatkov, S.G.; Soldatov, O.A.; Identification of the Heat Transfer Co-
efficient from Boundary Integral Data, Siberian Mathematical Journal,
2024, 65(4), 824–839.

[25] Pyatkov, S.; Pronkina, T. Coefficient Inverse Problems of Identification

of Thermophysical Parameters from Boundary Integral Data, Journal
of Mathematical Sciences, 282(2), 241–252.

[26] Triebel H. Interpolation Theory. Function Spaces. Differential Opera-

tors. VEB Deutscher Verlag der Wissenschaften: Berlin, 1978.

[27] Amann H. Compact embeddings of vector-valued sobolev and besov

spaces, Glasnik matematicki, 2000, 35(55), 161–177.

[28] Amann H. Linear and Quasilinear Parabolic Problems, Birkhauser Ver-

lag: Basel, 1995.

[29] Ladyzhenskaya, O.A.; Solonnikov, V.A.; Uraltseva, N.N. Linear and

Quasilinear Equations of Parabolic Type; American Mathematical So-
ciety: Providence, RI, USA, 1968.

[30] Baranchuk, V.A.; Pyatkov, S.G. On some classes of inverse problems

with pointwise ovedetermination for mathematical models of heat and
mass transfer. Bulletin of the Yugra State University, 2020, 3, 38–48.

14
Sergey Pyatkov, Doctor of Sci., Professor, Engineering School of Digital
Technologies, Yugra State University, Khanty-Mansiysk, Russia, E-mail: py-
[email protected];
Oleg Soldatov, postraduate, Engineering School of Digital Technologies, Yu-
gra State University, Khanty-Mansiysk, Russia, E-mail: [email protected]

15