An Egyptian fraction is a sum of positive (usually) distinct unit fractions. The famous Rhind papyrus, dated to
around 1650 BC contains a table of representations of as Egyptian fractions for odd
between 5 and 101. The reason the Egyptians
chose this method for representing fractions is not clear, although André
Weil characterized the decision as "a wrong turn" (Hoffman 1998, pp. 153-154).
The unique fraction that the Egyptians did not represent using unit fractions was
2/3 (Wells 1986, p. 29).
Egyptian fractions are almost always required to exclude repeated terms, since representations such as
are trivial. Any rational number has representations
as an Egyptian fraction with arbitrarily many terms and with arbitrarily large denominators, although for a given fixed number of terms,
there are only finitely many. Fibonacci proved that any fraction can be represented as a sum of distinct unit fractions (Hoffman
1998, p. 154). An infinite chain of unit fractions can be constructed using
the identity
(1)
|
Martin (1999) showed that for every positive rational number, there exist Egyptian fractions whose largest denominator
is at most
and whose denominators form a positive proportion
of the integers up to
for sufficiently large
. Each fraction
with
odd has an Egyptian fraction
in which each denominator is odd
(Breusch 1954; Guy 1994, p. 160). Every
has a
-term representation where
(Vose 1985).
No algorithm is known for producing unit fraction representations having either a minimum number of terms or smallest possible denominator (Hoffman 1998, p. 155). However, there are a number of algorithms (including the binary remainder method, continued fraction unit fraction algorithm, generalized remainder method, greedy algorithm, reverse greedy algorithm, small multiple method, and splitting algorithm) for decomposing an arbitrary fraction into unit fractions. In 1202, Fibonacci published an algorithm for constructing unit fraction representations, and this algorithm was subsequently rediscovered by Sylvester (Hoffman 1998, p. 154; Martin 1999).
Taking the fractions 1/2, 1/3, 2/3, 1/4, 2/4, 3/4, ... (the numerators of which are OEIS A002260, and the denominators of which
are
copies of the integer
), the unit fraction representations using the greedy
algorithm are
(2)
| |||
(3)
| |||
(4)
| |||
(5)
| |||
(6)
| |||
(7)
| |||
(8)
| |||
(9)
| |||
(10)
| |||
(11)
|
The number of terms in these representations are 1, 1, 2, 1, 1, 2, 1, 2, 2, 3, 1, ... (OEIS A050205). The minimum denominators for each representation are given by 2, 3, 2, 4, 2, 2, 5, 3, 2, 2, 6, 3, 2, ... (OEIS A050206), and the maximum denominators are 2, 3, 6, 4, 2, 4, 5, 15, 10, 20, 6, 3, 2, ... (OEIS A050210).
The Egyptian fractions for various constants using the greedy algorithm are summarized in the following table.
constant | OEIS | Egyptian fraction
for |
A006487 | 3, 13, 253, 218201, 61323543802, ... | |
A118325 | 2, 5, 32, 1249, 5986000, 438522193400489, ... | |
A069139 | 2, 5, 141, 68575, 32089377154, ... | |
A006525 | 2, 5, 55, 9999, 3620211523, 25838201785967533906, ... | |
A006526 | 3, 29, 15786, 513429610, 339840390654894740, ... | |
A110820 | 2, 13, 3418, 52016149, 153922786652714666, ... | |
A118323 | 2, 3, 13, 176, 36543, ... | |
A117116 | 2, 9, 145, 37986, 2345721887, ... | |
A118324 | 2, 6, 38, 6071, 144715221, ... | |
A001466 | 8, 61, 5020, 128541455, 162924332716605980, ... | |
A006524 | 4, 15, 609, 845029, 1010073215739, ... |
Any fraction with odd denominator can be represented as a finite sum of unit fractions, each having an odd denominator (Starke 1952, Breusch 1954). Graham proved that infinitely many fractions with a certain range can be represented as a sum of units fractions with square denominators (Hoffman 1998, p. 156).
Paul Erdős and E. G. Straus have conjectured that the Diophantine equation
(12)
|
always can be solved, an assertion sometimes known as the Erdős-Straus conjecture, and Sierpiński (1956) conjectured that
(13)
|
can be solved (Guy 1994).
The harmonic number is never an integer except for
.
This result was proved in 1915 by Taeisinger, and the more general results that any
number of consecutive terms not necessarily starting with 1 never sum to an integer
was proved by Kürschák in 1918 (Hoffman 1998, p. 157). In 1932,
Erdős proved that the sum of the reciprocals of any number of equally spaced
integers is never a reciprocal.
Nontrivial sets of integers are known whose reciprocals sum to small integers. For example, there exists a set of 366 positive integers (with maximum 992) whose sum of reciprocals is exactly 2 (Mackenzie 1997; Martin). A similar set of 453 small positive integers is known that sums to 6 (Martin).