Integral of inverse functions
In mathematics, integrals of inverse functions can be The gure on the right is a proof without words of this
computed by means of a formula that expresses the formula. Laisant does not discuss the hypotheses neces-
antiderivatives of the inverse f 1 of a continuous and sary to make this proof rigorous, but it can be made ex-
invertible function f , in terms of f 1 and an antideriva- plicit with the help of the Darboux integral[2] (or Fubinis
tive of f . This formula was published in 1905 by theorem[3] if a demonstration based on the Lebesgue in-
Charles-Ange Laisant.[1] tegral is desired). Laisants third proof uses the addi-
tional hypothesis that f is dierentiable. Beginning with
f 1 (f (x)) = x , one multiplies by f (x) and integrates
1 Statement of the theorem both sides. The right-hand side is calculated using inte-
gration by parts to be xf (x) f (x) dx , and the for-
mula follows.
Let I1 and I2 be two intervals of R . Assume that f :
I1 I2 is a continuous and invertible function, and let Nevertheless, it can be shown that this theorem holds even
f 1 denote its inverse I2 I1 . Then f and f 1 have if f or f 1 is not dierentiable:[2][3] it suces, for ex-
antiderivatives, and if F is an antiderivative of f , the ample, to use the Stieltjes integral in the previous argu-
possible antiderivatives of f 1 are: ment. On the other hand, even though general mono-
tonic functions are dierentiable almost everywhere, the
proof of the general formula does not follow, unless f 1
f 1 (y) dy = yf 1 (y) F f 1 (y) + C, is absolutely continuous.[3]
It is also possible to check that for every y in I2 , the
where C is an arbitrary real number. derivative of the function y yf 1 (y) F (f 1 (y)) is
equal to f 1 (y) .[4] In other words:
(x + h)f (x + h) xf (x) (F (x + h) F (x))
x I1 lim = x.
h0 f (x + h) f (x)
bd To this end, it suces to apply the mean value theorem
d to F between x and x + h , taking into account that f is
monotonic.
2 Examples
1. Assume that f (x) = exp(x) , hence f 1 (y) =
ln(y)
. The formula above gives immediately
c ln(y) dy = y ln(y) y + C.
ac
2. Similarly, with f (x) = cos(x) and f 1 (y) =
a b arccos(y)
,
arccos(y) dy = y arccos(y)sin(arccos(y))+C.
Proof without words of the theorem 3. With f (x) = tan(x) and f 1 (y) = arctan(y)
,
In his 1905 article, Laisant gave three proofs. First, under arctan(y) dy = y arctan(y) +
the additional hypothesis that f 1 is dierentiable, one ln | cos(arctan(y))| + C.
may dierentiate the above formula, which completes the
proof immediately. His second proof was geometric. If
f (a) = c and f (b) = d , the theorem can be written:
3 History
d b Apparently, this theorem of integration was discovered
f 1 (y) dy + f (x) dx = bd ac. for the rst time in 1905 by Charles-Ange Laisant,[1]
c a
1
2 6 SEE ALSO
who could hardly believe that this theorem is new, and [5] Read online
hoped its use would henceforth spread out among stu-
[6] Parker, F. D. (JunJul 1955). Integrals of inverse func-
dents and teachers. This result was published indepen-
tions. The American Mathematical Monthly. 62 (6):
dently in 1912 by an Italian engineer, Alberto Caprilli, in 439440. doi:10.2307/2307006.
an opuscule entitled Nuove formole d'integrazione.[5] It
was rediscovered in 1955 by Parker,[6] and by a number [7] It is equally possible that some or all of them simply re-
of mathematicians following him.[7] Nevertheless, they called this result in their paper, without referring to pre-
all assume that f or f 1 is dierentiable. The general ver- vious authors.
sion of the theorem, free from this additional assumption,
[8] Michael Spivak, Calculus (1967), chap. 13, pp. 235.
was proposed by Michael Spivak in 1965, as an exercise
in the Calculus,[8] and a fairly complete proof following [9] See for instance the formula at the end of his article.
the same lines was published by Eric Key in 1994.[2] This
proof relies on the very denition of the Darboux inte- Staib, J. H. (Sep 1966). The Integration of Inverse
gral, and consists in showing that the upper Darboux sums Functions. Mathematics magazine. 39 (4): 223
of the function f are in 1-1 correspondence with the lower 224. doi:10.2307/2688087
Darboux sums of f 1 . In 2013, Michael Bensimhoun, es-
timating that the general theorem was still insuciently
known, gave two other proofs:[3] The second proof, based
on the Stieltjes integral and on its formulae of integration
6 See also
by parts and of homeomorphic change of variables, is the
most suitable to establish more complex formulae.[9] Young inequality
4 Generalization to holomorphic
functions
The above theorem generalizes in the obvious way to
holomorphic functions: Let U and V be two open and
simply connected sets of C , and assume that f : U V
is a biholomorphism. Then f and f 1 have antideriva-
tives, and if F is an antiderivative of f , the general an-
tiderivative of f 1 is
G(z) = zf 1 (z) F f 1 (z) + C.
Because all holomorphic functions are dierentiable, the
proof is immediate by complex dierentiation.
5 References
[1] Laisant, C.-A. (1905). Intgration des fonctions in-
verses. Nouvelles annales de mathmatiques, journal des
candidats aux coles polytechnique et normale. 5 (4): 253
257.
[2] Key, E. (Mar 1994). Disks, Shells, and Integrals of In-
verse Functions. The College Mathematics Journal. 25
(2): 136138. doi:10.2307/2687137.
[3] Bensimhoun, Michael (2013). On the antiderivative of
inverse functions. arXiv:1312.3839 .
[4] This very simple proof of the general theorem, the only
one that does not make use of integrals, was communi-
cated by the French mathematician and Wikipedian Anne
Bauval in the corresponding pages in French. It seems to
have escaped the persons who published proofs of this re-
sult.
3
7 Text and image sources, contributors, and licenses
7.1 Text
Integral of inverse functions Source: https://en.wikipedia.org/wiki/Integral_of_inverse_functions?oldid=743450036 Contributors:
Michael Hardy, Headbomb, Rankersbo, Ozob, Yobot, Citation bot, LilHelpa, Trappist the monk, D.Lazard, Maschen, BG19bot, BattyBot,
Sgruenwald, Mark viking, Anne Delong, Maimonid, Gzer, Ktmrxit, Monkbot and Anonymous: 5
7.2 Images
File:FunktionUmkehrIntegral2.svg Source: https://upload.wikimedia.org/wikipedia/commons/5/59/FunktionUmkehrIntegral2.svg Li-
cense: CC BY-SA 3.0 Contributors: Own work Original artist: Jonathan Steinbuch
File:Lock-green.svg Source: https://upload.wikimedia.org/wikipedia/commons/6/65/Lock-green.svg License: CC0 Contributors: en:File:
Free-to-read_lock_75.svg Original artist: User:Trappist the monk
7.3 Content license
Creative Commons Attribution-Share Alike 3.0
