(Analysis by Nick Wu)

There are two parts to solving this problem. Firstly, we need to be able to efficiently verify whether a given value of $K$ is valid. Secondly, we need to efficiently find the minimum value of $K$ that is valid.

We start by determining whether a given value of $K$ is valid. We need to simulate the process of having cows dance on the stage and then leaving the stage. Whenever $K$ cows are on the stage and we want to figure out when the next cow can join, we want to know the earliest time when a cow can leave the stage. This motivates the idea of using a priority queue, which supports efficient insertion as well as efficient removal of the minimum element stored within (both in $O(\log K)$ time). The process of checking if a given value of $K$ is correct therefore takes $O(N \log K)$ time.

It remains to efficiently determine the minimum value of $K$ that is valid. Because $N$ is at most $10^4$, it is possible to iterate $K$ from $1$ to $N$ to find such a value for an $O(N^2 \log N)$ algorithm. We can do better though!

We are guaranteed that $K=N$ is a valid value. Given the structure of the problem, note that if $K$ is a valid value, then $K+1$ is a valid value - this is because having an extra spot cannot make cows dance later than they originally would. Therefore, we can binary search for the minimum valid value of $K$, giving us an $O(N \log^2 N)$ algorithm.

import java.io.*;
import java.util.*;
public class cowdance {
	public static void main(String[] args) throws IOException {
		BufferedReader br = new BufferedReader(new FileReader("cowdance.in"));
		PrintWriter pw = new PrintWriter(new BufferedWriter(new FileWriter("cowdance.out")));
		StringTokenizer st = new StringTokenizer(br.readLine());
		int n = Integer.parseInt(st.nextToken());
		int maxT = Integer.parseInt(st.nextToken());
		int[] l = new int[n];
		for(int i = 0; i < n; i++) {
			l[i] = Integer.parseInt(br.readLine());
		}
		int min = 1;
		int max = n;
		while(min != max) {
			int mid = (min+max)/2;
			if(possible(l, mid, maxT)) {
				max = mid;
			}
			else {
				min = mid+1;
			}
		}
		pw.println(min);
		pw.close();
	}
	
	public static boolean possible(int[] l, int k, int t) {
		int lastTime = 0;
		PriorityQueue<Integer> q = new PriorityQueue<Integer>();
		for(int i = 0; i < l.length; i++) {
			if(q.size() == k) {
				lastTime = q.poll();
			}
			if(lastTime + l[i] > t) {
				return false;
			}
			q.add(lastTime + l[i]);
		}
		return true;
	}
	
}