Consider a table of 7 rows and 7 columns. We want to fill some of the cells with a , such that the intersection of any 3 (not necessarily consecutive) rows and any 3 (not necessarily consecutive) columns will contain a cell with a . What is the minimum number of required?
Details and assumptions
Clarification: The 4 by 4 table with only 1 in the cell doesn't satisfy the conditions, because the intersection of rows 1, 2, 4, and columns 1, 3, 4, do not contain any cell with a .
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
First, we have to show that 15 ∗ isn't sufficient, i.e. that we can find 3 rows and columns which do not have a ∗ in them. Proof by contradiction. Suppose such a configuration exists. If there exists 3 rows that contain at most 4 ∗ , then we can find 3 columns which do not have any ∗ in these 3 rows to give the contradiction. Consider the number of rows with at most 2 ∗ . If there are at most 4 such rows, then there are 3 rows which contain at least 9 ∗ , so the other 4 rows contain at most 6 ∗ . By the pigeonhole principle, there are 3 rows that contain at most ⌊ 4 6 × 3 ⌋ = 4 ∗ , so we are done. If there are at least 5 such rows, they will contain at most 10 ∗ . By the pigeonhole principle, there are 3 rows which contain at most ⌊ 7 1 0 × 3 ⌋ = 4 ∗ . Hence, we reach a contradiction again.
It remains to show that 16 ∗ is sufficient. We can check that it is, using ( 1 , 1 ) , ( 2 , 2 ) , ( 2 , 3 ) , ( 3 , 4 ) , ( 3 , 5 ) , ( 4 , 6 ) , ( 4 , 7 ) , ( 5 , 2 ) , ( 5 , 4 ) , ( 5 , 6 ) , ( 6 , 2 ) , ( 6 , 5 ) , ( 6 , 7 ) , ( 7 , 2 ) , ( 7 , 3 ) , ( 7 , 7 ) .