 ANUARM - Editorial

#1

Author: Anudeep Nekkanti
Tester: Constantine Sokol
Editorialist: Florin Chirica

Simple

none

PROBLEM:

You have M vectors of length N. Each vector is defined by a position pos. Array on position i is pos – i if i <= pos and i – pos if i > pos. For each position, find maximum of values of all M vectors.

QUICK EXPLANATION

It turns out it’s enough to calculate minimal and maximal value of all M positions. For an index i, the answer is max(|i – minimal_position|, |i – maximal_position|), where |x| is absolute value of x.

Special case of M = 1

A good start is to see what happens if M = 1. Suppose selected position by the captain is pos. Then,

• numbering* = 0 if i = pos
• numbering* = numbering[i – 1] + 1 if i > pos
• numbering* = numbering[i + 1] + 1 if i < pos

where numbering is the resulting array. We can see that if i > pos, numbering* = i – pos and if i < pos, numbering* = pos – i. But this is the definition of absolute value. So numbering* = |i – pos| (where by |x| I denote absolute value of x).

Take general case

For general case, suppose pos, pos, …, pos[M] are positions asked by the captain during the M rounds. For each i (0 <= i < N) you need to output max(|i – pos|, |i – pos|, …, |i – pos[M]|).
So now let’s calculate for a particular i (let’s forget for a moment we need to iterate each i, we’re interested only in a single value of i). Let’s maximize |i – a|, where a can be pos, pos, …, pos[M]. A property of absolute value says that |i – a| = max(a – i, i – a). So if we maximize the expressions (a – i) and (i – a), then take their maximum value, we’re done for the given i.

In both expressions i is a constant (since we’ve decided to calculate only for a particular value of i). Since i is constant, it can’t influence maximization. So we need to maximize (a) and (–a). Now it’s obvious we need to take maximum and minimum between pos, pos, …, pos[M]. Let max_value be the maximum and min_value be the minimum. The answer for a fixed i is max(max_value – i, i – min_value).

The only remained thing is to iterate over all possible i. We calculate at the beginning max_value and min_value, then we iterate i and print max(max_value – i, i – min_value).

Time Complexity

The algorithm solves each test in O(N + M) time complexity.

AUTHOR’S AND TESTER’S SOLUTIONS:

#2

where is the practice link? Btw nice question. #3

What is wrong with the below approach?
For every M in the input calculate the maximum value of the left most guy(index 0)[leftMax] and right most guy(index n-1)[rightMax]
At the end the maximum value assigned to guy at position i is maximum(|leftMax-i|,|rightMax-(n-1-i)|)?

#4

http://www.codechef.com/viewsolution/6397057

#5

https://www.codechef.com/viewsolution/14065014

#6

Good question,btw i am unable to comment! how do i gain upvotes when i can neither write a comment or pose a question?

#7

Please can you provide what is output to this input
1
6 3
2 3 0

#8

Getting access denied errors in XML when I open the Author’s and Tester’s solutions.

#9

What has the tester done in so much code?

#10

#11

You are not printing a space after each number

#12

3 2 2 3 4 5