The Egyptian fraction is a sum of unique fractions with a unit numerator (unit fractions). There is an infinite number of ways to represent a fraction as a sum of unit fractions. Several methods have been developed to convert a fraction to this form. This calculator can be used to expand a fractional number to an Egyptian fraction using Splitting, Golomb, Fibonacci/Sylvester, Binary, or Bleicher/Erdős methods1. Enter any number between 0 and 1 in decimal or simple fraction form, and the calculator will expand it to a sum of distinct unit fractions. The calculator also can try to find the best method among listed above, minimizing either denominators sum or maximal denominator (see below for more details on best method criteria).
See Egyptian Fraction to Rational Number Converter for the inverse transformation.
Ancient Egyptians did not use the fraction expansion methods mentioned above to represent a fraction as a unit fraction sum. We can consider that analyzing ancient documents surviving to this day. The calculator below uses the algorithms mentioned above to expand fractions with the numerator 2 and the odd denominator in the range from 5 to 101 and compare the results with the Rhind papyrus (1650 B.C.E). The Golomb method does not participate in comparison since it gives the same results as the Fibonacci/Sylvester method for the Rhind papyrus data (and all fractions with numerator = 2 in the general case).
The following comparison criteria give the best results for original data from Rhind papyrus comparing to all method results:
Both comparison criteria chooses the Rhind papyrus fraction expansions as the best in 46 of 49 cases. The Fibonacci/Sylvester methods wins for Minimise : Hieroglyph count and Minimise : Terms count criteria. But if you slightly change the method of counting hieroglyphs (if you count as one any dash set, denoting numbers from 2 to 9), the Rhind papyrus expansion will only lose slightly to the Fibonacci/Sylvester method.
Kevin Gong, Egyptian Fractions, UC Berkeley Math 196 Spring 1992 ↩