In an M x N matrix all non-zero entries are covered in 'a' rows and 'b' columns. Then the maximum number of non-zero entries, such that no two are on the same row or column, is:
A. a <= b
B. <= min(a,b)
C. <= a + b
D. <= min(M-a, N-b)
GATE 2004 · Engineering Mathematics · Matrix · medium
Answer: The maximum number is <= min(M-a, N-b). Answer: D.
Identify where non-zero entries can be: Any non-zero entry must be in one of the a covered rows AND one of the b covered columns. The (M-a) uncovered rows and (N-b) uncovered columns contain only zeros.
Apply the bound for independent selection: By König's theorem: the maximum matching in a bipartite graph equals the minimum vertex cover. Here, the minimum cover has a+b vertices. The maximum independent set (no two in same row or column from non-zero entries) is bounded by min(M-a, N-b).