Volume 25, N. 2-3, pp.

133138, 2006
Copyright 2006 SBMAC
ISSN 0101-8205
www.scielo.br/cam
On an inverse boundary value problem*
ALBERTO P. CALDERN
Department of Mathematics, The University of Chicago
Abstract. This paper is a reprint of the original work by A. P. Caldern published by the
Brazilian Mathematical Society (SBM) in ATAS of SBM (Rio de Janeiro), pp. 65-73, 1980. The
original paper had no abstract, so this reprint to be truthful to the original work is published with
no abstract.
Mathematical subject classication: 35J25, 35Q60, 47F05, 86A20, 86A22.
Key words: inverse problems, boundary value problems, identication problem, elliptic equa-
tions.
In this note we shall discuss the following problem. Let D be a bounded do-
main in R
n
, n 2, with Lipschitzian boundary dD, and be a real bounded
measurable function in D with a positive lower bound. Consider the differential
operator
L

(w) = ( w)
acting on function of H
1
(D) and the quadratic form Q

() , where the functions

in H
1
(R
n
) , dened by
Q

() =
_
D
(w)
2
dx, w H
1
_
R
n
_
, w|
d D
= 0
L

w = ( w) = 0 in D.
#693/06. Received: 14/XI/06.
*We thank the Brazilian Mathematical Society (SBM) for the permission to reproduce
this article originally published at the Seminar on Numerical Analysis and its Application to
Continuum Physics in the ATAS of SBM (Rio de Janeiro), 1980, pp. 6573.
134 ON AN INVERSE BOUNDARY VALUE PROBLEM
The problem is then to decide whether is uniquely determined by Q

and to
calculate in terms Q

, if is indeed determined by Q

.
This problem originates in the following problem of electrical prospection. If
D represents an inhomogeneous conducting body with electrical conductivity
, determine by means of direct current steady state electrical measurements
carried out on the surface of D, that is without penetrating D. In this physical sit-
uation Q

() represents the power necessary to maintain an electrical potential

in D.
In principle Q

can be determined through measurements effected on d D and

contains all information about which can be thus obtained.
But let us return to our mathematical problem. Let us introduce the following
norms in the space of functions on dD and in the space of quadratic forms
Q()

2
=
_
D
|u|
2
dx ; u|
d D
= , u = 0 in D.
Q = sup
1
|Q()|
Then the mapping
: Q

is bounded and analytic in the subset of L

(D) consisting of functions which

are real and have a positive lower bound. Our goal is then to determine whether
is injective, and invert if this is the case. This we are yet unable to do and
is, as far as we know, an open problem. However we shall show that d|
=const
is indeed injective, that is, the linearized problem has an afrmative answer. If
d|
=const
, which is a linear operator, had a closer range, one could conclude
that itself is injective in a sufciently small neighborhood of = const. But
the range of d is not closed and the desired conclusion cannot be obtained in
this fashion. Nevertheless, as we shall see below, if is sufciently close to a
constant, it is nearly determined by Q

and in some cases it can be calculated

with an error much smaller than const
L
.
To show this let us rst obtain an expression for the solution of the equation
L

(W) = ( W) , W|
d D
= . = 1 +
Comp. Appl. Math., Vol. 25, N. 2-3, 2006
ALBERTO P. CALDERN 135
Let W = u + v, where u = L
1
u = 0, u|
d D
= . Then
L

(W) = L
1+
(u + v) = L
1
v + L

v + L

u = 0
Since u|
dD
= W|dD we have v|
d D
= 0 and v H
1
0
(D), the closure in H
1
(R
n
)
of functions of C

with support in D. But L

1
, as an operator from H
1
0
(D) into
H
1
(D), has a bounded inverse G, and from the last expression we obtain
v + GL

v = GL

u.
and
v =
_

0
(1)
j
(GL

)
j
_
(GL

u) (1)
Since for W H
1
0
(D), L

W
H
1
(D)

L
W
H
1
0
(D)
if A denotes the
norm of G, the series above will converge for
L
A < 1 and
v
H
1
(D)

A
L

1 A
L

(2)
From (1) it follows that is analytic at = 1. The same argument would show
that is analytic at any other point .
Next let us calculate d|
=1
. We have
Q
1+
() =
_
D
(1 + )|W|
2
dx =
_
D
[(1 + )|u|
2
+ 2(u u)
+ 2(u v) + (1 + )|v|
2
]dx
(3)
The contribution of the second term in the integrand of the last integral vanishes
on account of the fact that u = 0. Furthermore, from (1) one sees readily that
the parts linear in of the last two terms in the integrand vanish. Thus setting
d = we obtain
dQ

()|
=1
=
_
D
|u|
2
dx, u = 0, u|
d D
=
To show that dQ

()|
=1
is injective, we merely have to show that if the last
integral vanishes for all u with u = 0 then = 0. But if the integral vanishes
for all such u, then we also have
_
D
(u
1
.u
2
)dx = 0 (4)
Comp. Appl. Math., Vol. 25, N. 2-3, 2006
136 ON AN INVERSE BOUNDARY VALUE PROBLEM
whenever u
1
= u
2
= 0 in D. Now let Z be any vector in R
n
and a another
vector such that |a| = |Z|, ( a.Z) = 0. Then the functions
u
1
(x) = e
i (Z.x)+(a.x)
, u
2
= e
i (Z.x)(a.x)
. (5)
are harmonic, and substituting in (3) we obtain
2|Z|
2
_
D
(x)e
2i (Z.x)
dx = 0, Z
whence it follows that = 0.
Nowlet us return to Q

(W). We set again = 1+ and introduce the bilinear

form
B(
1
,
2
) =
1
2
_
Q

(W
1
+ W
2
) Q

(W
1
) Q

(W
2
)
_
and setting W
j
= u
j
+ v
j
, j = 1, 2, u
j
= 0, u
j
|
d D
=
j
we obtain
B(
1
,
2
) =
_
D
(1 + )(u
1
.u
2
) +
_
(u
1
.v
2
) + (u
2
.v
1
)
_
+ (1 + )(v
1
.v
2
)dx
Now, substitution of the exponentials in (5) for u
1
and u
2
in the preceding
expression (taking a to be a function of Z such that |a| = |Z|, (a.Z) = 0) yields
(Z) =

F(Z) + R(Z) (6)
where (Z) is a Fourier transform of extended to be zero outside D, the
function

F(Z) =
1
2
2
|Z|
2
B
_
e
i (Z.x)+(a.x)
, e
i (Z.x)(a.x)
_
is known and, as follows readily from (2),
|R(Z)| C
2
L
e
2r|Z|
(7)
provided that A
L
1 , where C depends only on D and , and r is the
radius of the smallest sphere containing D. Now R(Z) is too large to permit
estimating (x). However, under favorable circumstances it is still possible to
obtain satisfactory information about . Choose , 1 < < 2, then for
|Z|
2
2r
log
1

= (8)
Comp. Appl. Math., Vol. 25, N. 2-3, 2006
ALBERTO P. CALDERN 137
we have |R(Z)| C

L
. Let be a function such that C

, supp
{|Z| 1} (0) = 1, and let

=
n
( Z). Then we have
(Z)
_
Z

_
=

F(Z)
_
Z

_
+ R(Z)
_
Z

_
and
(

)(x) = (F

)(x) + (x) (9)

where denotes convolution and
|(Z)| C

(
Z

dZ = C
1

_
log
1

_
n
where C
1
depends only on D, e .
Thus if
L
is sufciently small, (9) gives an approximation for

with
an error which is much smaller than
L
. Clearly, if
L
is small then is
large and

is itself in some sense, a good approximation to .

Approximations to the function itself may be obtained if one assumes that
, extended to be equal to 1 outside D, is in C
m
, m > n. In this case one obtains

(Z) =

F
1
(Z) +

R (Z)
where F
1
is known and R (Z) is the same as in (6). One then calculates (x) by
integrating over |Z| with as in (8) and estimates the error by using the
decay of

at . Thus one obtains
(x) = F
2
(x) + (x)
where F
2
(x) is known and
| (x)| C

_
log
1

_
n
+ CM
_
log
1

_
m+n
where M is a bound for the derivatives of order m of .
REFERENCES
We have been unable to nd treatments of the problem discussed above in the literature,
at least not in the general setting in which we are interested. Similar problems have been
studied in the papers listed below.
Comp. Appl. Math., Vol. 25, N. 2-3, 2006
138 ON AN INVERSE BOUNDARY VALUE PROBLEM
[1] J.R. Cannon, Determination of certain parameters in heat conduction problems. J. Math.
Anal. Appl., 8 (1964), 188201.
[2] J.R. Cannon, Determination of an unknown coefcient in a parabolic differential equation.
Duke Math. J., 30 (1963), 313323.
[3] J.R. Cannon, Determination of the unknown coefcient k (u) in the equation V k (u) u = 0
from overspecied boundary data. J. Math. Anal. Appl., 18 (1967), 112114.
[4] J.R.Cannon and Paul Du Chateau, Determining unknown coefcients in a non-linear heat
conduction problem. SIAM J. Appl. Math., 24(3) (1973), 298314.
[5] Jim Douglas, Jr and G.F. Jones, Determination of a coefcient in a parabolic differential
equation, Part. II: numerical approximation. J. Math. Mech., 11 (1962), 919926.
[6] B.F. Jones, Various methods for nding unknown coefcients in parabolic differential equa-
tions. Comm. Pure Appl. Math., 16 (1963), 3344.
Comp. Appl. Math., Vol. 25, N. 2-3, 2006