Question

我正在尝试通过查找矩阵中子矩形的最大总和来解决问题。到目前为止，我已经设法实现了大约90％的算法。（对于任何不知道我在说什么的人：https://www.youtube.com/watch?v=yCQN096CwWM）

我的问题是我不知道如何删除临时数组的顶部或底部值，如果它们是负数的话。更不用说如何知道删除的负值来自哪个索引。

我已经解决了1维（1D）中的问题：

int maxSoFar = 0;
int maxEndingHere = 0;
    int i;
    for (i = 0; i < input.length; i++) {
        System.out.println(i + " + " + input[i]);
        if (maxEndingHere + input[i] > 0) {
            maxEndingHere += input[i];
        } else {
            maxEndingHere = 0;
        }
        if (maxSoFar < maxEndingHere) {
            maxSoFar = maxEndingHere;
        }
    }
    System.out.println(maxSoFar);

在二维（2D）中，我已经走到了这一步：

    int[][] input = new int[4][5];
    input[0][0] = 2;
    input[0][1] = 1;
    input[0][2] = -3;
    input[0][3] = -4;
    input[0][4] = 5;
    input[1][0] = 0;
    input[1][1] = 6;
    input[1][2] = 3;
    input[1][3] = 4;
    input[1][4] = 1;
    input[2][0] = 2;
    input[2][1] = -2;
    input[2][2] = -1;
    input[2][3] = 4;
    input[2][4] = -5;
    input[3][0] = -3;
    input[3][1] = 3;
    input[3][2] = 1;
    input[3][3] = 0;
    input[3][4] = 3;
    int currentSum = 0;
    int maxSum = 0;
    int L;
    int R;
    int maxR = 0;
    int maxL = 0;
    int maxUp = 0;
    int maxDown = 0;
    int k;
    int j;
    int[][] temp = new int[4][1];

    for (L = 0; L < input.length; L++) {
        temp[0][0] = 0;
        temp[1][0] = 0;
        temp[2][0] = 0;
        temp[3][0] = 0;
        for (R = L; R < input[0].length; R++) {
            currentSum = 0;
            for (k = 0; k < temp.length; k++) {
                temp[k][0] += input[k][R];
                if (currentSum + input[k][R]<0){
                    currentSum += input[k][R];
                }
                else {currentSum = 0;}
                if (k == 3 && currentSum >= maxSum) {
                    maxSum = currentSum;
                    maxR = R;
                    maxL = L;
                }
            }
        }
    }
    System.out.println("Max sum = " + maxSum);
    System.out.println("Max R = " + maxR);
    System.out.println("Max L = " + maxL);

这是我在这里发表的第一篇文章，如果我遗漏了某些内容，请告诉我。

感谢您的协助！

Answer 1

我设法解决了问题，这是工作代码：

import java.util.Arrays;
public class Algoritme4 {
public static void main(String[] args) {
    algo2D(new int[][]{
            {2, 1, -3, -4, 5},
            {0, 6, 3, 4, 1},
            {2, -2, -1, 4, -5},
            {-3, 3, 1, 0, 3}
    });
}

private static int[] algo4(int[] a) {
    int maxEndingHere = 0;
    int i;
    int[] result = {0, 0, 0};
    for (i = 0; i < a.length; i++) {
        if (maxEndingHere + a[i] > 0) {
            maxEndingHere += a[i];
        } else {
            maxEndingHere = 0;
            result[1] = i + 1;
        }
        if (result[0] < maxEndingHere) {
            result[0] = maxEndingHere;
            result[2] = i;
        }
    }
    System.out.println("Max sum of sub matrix = " + result[0] + " (" + result[1] + "," + result[2] + ")");
    return result;
}
private static void algo2D(int[][] a){
    int[] result = {0,0,0};
    int maxSoFar = 0;
    int maxR = 0;
    int maxL = 0;
    int start = 0;
    int slut = 0;
    int L;
    int R;
    int k;

    for (L = 0; L<a.length; L++) {
        int[] temp = {0,0,0,0};
        for (R = L; R < a[0].length; R++){
            System.out.println("R = " + R);
            System.out.println("L = " + L);
            for (k = 0; k < a.length; k++) {
                temp[k] += a[k][R];
                if (k==3) {
                    System.out.println("temp = " + Arrays.toString(temp));
                    result = algo4(temp);
                }
                if (result[0]> maxSoFar){
                    maxSoFar = result[0];
                    maxL = L;
                    maxR = R;
                    start = result[1];
                    slut = result[2];
                }

            }

        }
    }
    System.out.println("Max: " + maxSoFar + " (" + maxL + "," + start + ") (" + maxR + "," + slut + ")");
}

}

矩阵中子半径的最大和（2D）

1 个答案: