pizzeria_report.txt

First thought was to go through each square and find number of available pizzerias.
However this would be N^2M complexity

Second idea is to construct an NxN map to keep track of number of available pizzerias.
Then just cycle through M, before going over them all to find the maximum.
Hence would be (MR^2 + N^2) complexity, R is small so significant improvement.

A minor change is to track the maximum as you go and so not have to do the N^2 scan at the end.

Complexity is O(N^2) memory, and O(N^2 + MR^2) for time. (N^2 from array initialization)


Alternative approach:
Instead of storing data in an array use a dictionary with location as key.
This could save on the O(N^2) time to initialize array.
This is particularly helpful if R is small.
I've left this method in comments.

complexity O(MR^2) which is better.
But I assume the worse cases will have a fairly saturated map of pizzerias and so the whole map will be used.
In this case the constants for the array are probably better.