Sums of unit fractions
A unit fraction is a fraction with 1 as its numerator, such as , or . Scribes in ancient Egypt wrote fractions as sums of different unit fractions. In that notation, is written .
The unit fractions in the sum must all be different. is also , but that repeats , so it is not allowed. A sum of different unit fractions is called an Egyptian fraction.
7 twelfths are shaded in the top bar. A half is 6 twelfths, and 1 twelfth more makes 7: .
Take the biggest unit fraction that fits
The greedy method finds an Egyptian fraction for any fraction between 0 and 1. Take the biggest unit fraction that fits, meaning the biggest one that is not more than the fraction. Subtract it. Then do the same to what is left, written in lowest terms, and stop when what is left is a unit fraction.
To find the biggest unit fraction that fits inside , divide b by a and, if the answer is not a whole number, round it up. This works because fits exactly when , which means n is at least , and the smallest such n gives the biggest unit fraction. For , 12 ÷ 7 = 1.7…, which rounds up to 2, so the biggest unit fraction that fits is . Subtract it: . That is a unit fraction, so the method stops: .
Now . 7 ÷ 5 = 1.4, which rounds up to 2, so take . . Next, 14 ÷ 3 = 4.6…, which rounds up to 5, so take . , a unit fraction. So .
numerator = 5, and every greedy step has to leave it smaller
Keep taking the biggest unit fraction that fits, until nothing is left over.
The bar is one whole, and the line marks . Take one greedy step at a time: , then , then . The numerator of what is left goes 5, 3, 1. Choose and its numerator goes 4, 3, 1.
Why the method always stops
In , the numerator of what was left went 5, then 3, then 1. It gets smaller at every step, for every fraction. Here is why.
Say what is left is , and the biggest unit fraction that fits is . The next unit fraction up, , did not fit, so it is more than , and b is more than a × (n − 1). Now subtract: . The new numerator is an − b = a − (b − a × (n − 1)). The bracket is at least 1, so the new numerator is less than a. It is not 0 either: that would make exactly , a unit fraction, and the method would already have stopped.
For , n = 2: the new numerator is 5 − (7 − 5 × 1) = 5 − 2 = 3, as found above. Canceling down afterward can only make the numerator smaller still.
A numerator is a whole number, and a whole number cannot get smaller forever. Starting from a, it reaches 1 after at most a − 1 steps. A numerator of 1 means that what is left is a unit fraction, and the method stops.
No unit fraction is used twice
What is left after taking is less than , so the next unit fraction taken is smaller than . The reason: is less than , and is at most , because 2(n − 1) is at least n whenever n is 2 or more. So is less than . Each unit fraction is smaller than the one before, so they are all different.
After , what is left of is , less than . After , what is left is , less than .
Canceling down along the way
Take . 13 ÷ 4 = 3.25 rounds up to 4, and . Next, 52 ÷ 3 = 17.3… rounds up to 18, and , which cancels down to . So , and the numerators went 4, 3, 1.
Greedy is not always shortest
The greedy method always finishes, but it does not always find the shortest sum or the smallest denominators. For it needs five unit fractions, and the third is already ,309. A shorter answer exists: . Check it over the common denominator 363: .
Worked example: Sharing Loaves Among Workers the Egyptian Way
Question In ancient Egypt a share was written as a sum of different unit fractions, which are fractions with a numerator of 1. The greedy method takes the largest unit fraction that is not more than the share, subtracts it, and repeats with what is left. (a) 5 loaves are shared equally among 8 workers. Write each worker's share as a sum of different unit fractions by the greedy method, and say what each worker receives. (b) On another day 11 loaves are shared equally among 12 workers. Write each share in the same way, and say how the 11 loaves are cut.
1.Each worker's share is 5 ÷ 8 = 58 of a loaf. A whole loaf is more than 58, and 12 = 48 is not more than 58, so the largest unit fraction that fits is 12.
Each share is 58 of a loaf, and the largest unit fraction that fits is 12 = 48. 2.Subtract it: 58 − 48 = 18. What is left is a unit fraction, so the method stops, and 58 = 12 + 18.
58 − 48 = 18, which is a unit fraction, so 58 = 12 + 18. 3.(a) Each worker receives half a loaf and one eighth of a loaf. Cut 4 loaves into halves to make 8 halves, and cut the fifth loaf into 8 eighths. Check: 4 + 1 = 5 loaves.
(a) 4 loaves are cut into 8 halves and 1 loaf into 8 eighths. Each worker receives a half and an eighth. 4.For 11 loaves among 12 workers the share is 1112. The largest unit fraction that fits is 12 = 612, and 1112 − 612 = 512 is left. Now 12 = 612 is too large, but 13 = 412 fits, and 512 − 412 = 112 is left.
For 1112: take 12 = 612, which leaves 512. Then take 13 = 412, which leaves 112. 5.(b) 1112 = 12 + 13 + 112. Cut 6 loaves into halves, 4 loaves into thirds and 1 loaf into twelfths, which makes 12 pieces of each size. Check: 6 + 4 + 1 = 11 loaves.
(b) 1112 = 12 + 13 + 112: 6 loaves in halves, 4 in thirds and 1 in twelfths.
Answer: (a) 58 = 12 + 18: each worker receives half a loaf and one eighth of a loaf; (b) 1112 = 12 + 13 + 112: 6 loaves are cut into halves, 4 loaves into thirds and 1 loaf into twelfths
Common mistakes
- Writing 58 = 18 + 18 + 18 + 18 + 18. The unit fractions must all be different. Repeating 18 also gives each worker 5 small pieces instead of 2 large ones.
- Taking 13 first for 1112 because it is a familiar fraction. The greedy method always takes the largest unit fraction that fits, and 12 fits, because 612 is not more than 1112.
The usual mistakes
Repeating a unit fraction. is true, but it is not an Egyptian fraction. The greedy method gives .
Not taking the biggest unit fraction that fits. For , fits, but so does , which is bigger. Greedy takes .
Taking a unit fraction that is too big. For , is more than , so it does not fit. 10 ÷ 3 = 3.3… rounds up to 4, and is the biggest that fits.