如何优化初学者HackerEarth代码

时间:2018-12-21 23:59:58

标签: java algorithm optimization data-structures

我一直在尝试解决HackerEarth问题,但最后几个测试用例似乎总是超时

问题陈述: https://www.hackerearth.com/practice/algorithms/searching/linear-search/practice-problems/algorithm/joker-and-thieves-53e59f4a/

我是一年级学生,所以我真的不知道该如何优化

我尝试查看java解决方案,但是它只是将大写的案例放入代码中,有效地消除了对其进行优化的需要

import java.io.*;
import java.util.*;
public class police {
    static int t,n,k;
    static char[][] test;
    static int max = 0;
    public static void main(String[] args) throws IOException {
        BufferedReader br = new BufferedReader(new InputStreamReader(System.in));
        StringTokenizer st = new StringTokenizer(br.readLine());
        t = Integer.parseInt(st.nextToken());


        for (int i = 0; i < t; i++) {
           st = new StringTokenizer(br.readLine());
            n = Integer.parseInt(st.nextToken());
            k = Integer.parseInt(st.nextToken());
            test = new char[n][n];


            int ret = 0;

            for (int b = 0; b < n; b++) {
                st = new StringTokenizer(br.readLine());
                for (int a = 0; a < n; a++) {
                    test[b][a] = st.nextToken().charAt(0);
                }
            }

            for (int b = 0; b < n; b++) {
                ret += solve(test[b]); //calculate each row
            }

            System.out.println(ret);


        }    


    }
    static int solve(char[] a) { //given a row, calculate the maximum number of catches
        int ret = 0;
        for (int i = 0; i < n; i++) {
            if (a[i] == 'P') {

                for (int b = i - k; b <= i + k; b++) { //scan area to see if police can catch a thief
                    if (b >= 0 && b < n && a[b] == 'T') {
                        a[b] = 'C'; //caught
                        break;
                    }
                }

                a[i] = 'U'; //used
            }

        }
        for (int i = 0; i < n; i++) //count
            if (a[i] == 'C')
                ret++;

        return ret;
    }
}

我很确定,如果有人可以帮助我,那将与解决方法有关

1 个答案:

答案 0 :(得分:2)

是的,在solve方法中使用的方法是造成超时的原因。

首先,您需要了解algorithm complexityBig O notation

问题约束是:

1 <= N <= 1000

1 <= K <= N * N

这些表明您的解决方案的复杂度最多应为O(N * N)。换句话说,最多有两个嵌套的for循环,每个循环的复杂度为O(N)。

在解决方案中,您将执行以下操作:

for (int b = 0; b < n; b++) {
    ret += solve(test[b]); //calculate each row
}

好的,此循环至关重要,因为您必须遍历网格中的所有行。复杂度:O(N)

然后在您的求解方法中:

for (int i = 0; i < n; i++) {
    if (a[i] == 'P') {

        for (int b = i - k; b <= i + k; b++) { //scan area to see if police can catch a thief
            if (b >= 0 && b < n && a[b] == 'T') {
                a[b] = 'C'; //caught
                break;
            }
        }

        a[i] = 'U'; //used
    }

}

这些嵌套的for循环是问题的真正原因,对于较高的O(N),外部循环的复杂度为O(N),内部循环的复杂度也可能为K。因此,三个for循环的总复杂度可能达到O(N * N * N),这肯定会超时。

这是我解决这个问题的方法,我只修改了solve方法:

static int solve(char[] a) { //given a row, calculate the maximum number of catches
        int ret = 0;
        ArrayDeque<Integer> policeMenQueue = new ArrayDeque<>(); // queue for holding positions of policemen
        ArrayDeque<Integer> thievesQueue = new ArrayDeque<>(); // queue for positions of thieves
        for (int i = 0; i < n; i++) {
            if(!policeMenQueue.isEmpty()) { // check if the leftmost policeman can catch a thief at current position (i)
                Integer mostLeftPoliceMan = policeMenQueue.getFirst();
                if(i - mostLeftPoliceMan > k) { // if he cannot then we must remove him as he will no longer be able to catch any thieves
                    policeMenQueue.removeFirst();
                }
            }
            if(!thievesQueue.isEmpty()) { // check if the leftmost thief can be caught be a policeman at current position (i)
                Integer mostLeftThief = thievesQueue.getFirst();
                if(i - mostLeftThief > k) { // if not then we must remove him as he will no longer be caught by any policemen
                    thievesQueue.removeFirst();
                }
            }
            if(a[i] == 'P') {
                if(!thievesQueue.isEmpty()) { // the leftmost thief can be caught by current policeman
                    ret++;
                    thievesQueue.removeFirst(); // ok this thief is caught, remove him
                } else {
                    policeMenQueue.addLast(i); // just add the policeman to keep his position in the queue
                }
            }
            if(a[i] == 'T') {
                if(!policeMenQueue.isEmpty()) { // the current thief can be caught by the leftmost policeman
                    ret++;
                    policeMenQueue.removeFirst(); // ok this policeman has already caught a thief (used), remove him
                } else {
                    thievesQueue.addLast(i); // just add the thief to keep his position in the queue
                }
            }
        }
        return ret;
    }

我的想法是:从左到右在每一行上循环播放,并根据问题指出:每个警察只能抓到一个小偷,我们希望最大程度地捕获被抓的小偷,因此,每个小偷都被抓到是对我们有利的最左边的警察(我们从左到右)。

例如考虑以下行:

P P T T

想象一下K = 2

我们赞成将3位置的小偷捕获到1位置的警察,因为该警察无法捕获4位置的盗贼,当然也不能捕获2位置的警察1}}在这种情况下可以同时捕获两个小偷,但是请记住,我们必须最大化捕获的小偷的数量,每个警察只能捕获一个小偷,因此我们希望每个小偷都由可以捕获他的最左边的警察捕获。 >

我的解决方案取决于queue数据结构(Java中的ArrayDeque),如果您不知道它是如何工作的或者我为什么使用它,请看这里:https://www.tutorialspoint.com/data_structures_algorithms/dsa_queue.htm