435. Non-overlapping Intervals

truongductri01

truongductri01

Posted on May 30, 2023

435. Non-overlapping Intervals

Problem: Non-overlapping Intervals

We want to utilize the greedy method to find the solution to this problem.

Usually, the easiest start for this kind of problem is to re-organize the intervals using the start and the end value. We can achieve this by sorting the intervals using the following rules:

  1. Assume that we have interval1 and interval2
  2. If interval1.start < interval2.start, interval1 will be smaller and vice versa
  3. If interval1.start == interval2.start, then which ever interval has a smaller value of .end will be the smaller one

So, the simple algorithm to be applied is:

  • Sort the intervals using the rules mentioned above
  • Traverse through the list of intervals and each step will consider 2 intervals, interval1 and interval2
  • If they are overlapped, remove which ever is longer, since this will prevent potential overlapping on the later ones.

Here is the code:

class Solution {
    public boolean isOverlap(int[] interval1, int[] interval2) {
        return interval2[0] < interval1[1];
    }

    public boolean isShorter(int[] interval1, int[] interval2) {
        return interval1[1] <= interval2[1];
    }

    public int eraseOverlapIntervals(int[][] intervals) {
        /**
        greedy?

        can we immediately tell which intervals to be removed?
        can we re-arrange the intervals?

        [[1,2],[2,3],[3,4],[1,3]]

        sort based on primary the start and secondary the end date:
        [[1,2], [1,3], [2,3], [3,4]]

        check if the 2 intervals overlap, keep the one that is shorter, remove the one that is larger

        We can use the two pointers to keep track of which to be removed or which not.
        - first and second pointer
        - while second is withint the length of the intervals, continue
        - if the interval of the first and second pointer overlap:
            - increase the count
            - always remove the second since we sort based on both start and end date? No, we want the shorter intervals
            - check which one is shorter. If the first one is shorter, move second => second + 1
            - else, first = second, second = second + 1.
        - otherwise, first = second, second = second + 1. 

        - Run time: O(n log n) since we need to sort
        */

        Arrays.sort(intervals, (a, b) -> a[0] == b[0]? a[1] - b[1]: a[0] - b[0]);

        // print the result
        // System.out.println(Arrays.deepToString(intervals));

        if (intervals.length < 2) {
            return 0;
        }

        int count = 0;
        int first = 0;
        int second = 1;

        while (second < intervals.length) {
            int[] interval1 = intervals[first];
            int[] interval2 = intervals[second];

            if (isOverlap(interval1, interval2)) {
                count ++;
                // System.out.println(Arrays.toString(interval1) + " overlap " + Arrays.toString(interval2));

                if (isShorter(interval1, interval2)) {
                    // delete the second one 
                    second ++;
                } else {
                    first = second;
                    second ++;
                }
            } else {
                first = second;
                second ++;
            }
        }

        return count;
    }
}
Enter fullscreen mode Exit fullscreen mode
💖 💪 🙅 🚩
truongductri01
truongductri01

Posted on May 30, 2023

Join Our Newsletter. No Spam, Only the good stuff.

Sign up to receive the latest update from our blog.

Related