华为诺亚方舟实验室-CodingTest
Huawei - CodingTest
01 - 编程题01
A为一个十进制整数,k位,k<100。求B使得B为大于A的最小整数,且A各位的和等于B各位的和。请给出函数实现,测试例,并说明解题思路。
思路:
首先,以暴力算法扫描的话时间复杂度过大,我们根据规律可以发现:
从A的倒数第二位开始,从后向前搜索,直到找到一个符合条件的位数
A[i]
,使其满足如下条件:A[i]
不等于9并且A[i]
之后的所有位不全为0
这样,我们将满足以上要求的数据,按一下的策略进行修改:
A[i]
加1, 计算后面的所有位数之和记作sum
, 并从最后一位向前,将位数和减去9
并改为9
如果剩余无法补
9
则继续改为sum
若为0则补0
直到修改到A[i]
此外 还会有不满足以上要求的数字,则我们在开头先补上一个1
,再按我们的规律进行修改;
测试样例及对拍代码:
对拍代码即暴力解 时间复杂度高 以此来验证代码的正确性;
对拍代码(matcher.cpp):
/*
* @Description:
* @Versions:
* @Author: Vernon Cui
* @Github: https://github.com/vernon97
* @Date: 2021-03-26 18:48:04
* @LastEditors: Vernon Cui
* @LastEditTime: 2021-03-22 21:23:26
* @FilePath: /Projects/matcher.cpp
*/
#include <cstdlib>
#include <ctime>
#include <iostream>
const int N = 100;
const std::string bf = "../bf.cpp";
const std::string algo = "../Example.cpp";
/*
封装一下 `system`,支持 ctrl + c 退出整个对拍程序
*/
int mySystem(const char* command) {
int result = system(command);
if (WIFEXITED(result)) {
// printf("Exited normally with status %d\n", WEXITSTATUS(result));
} else if (WIFSIGNALED(result)) {
printf("Exited with signal %d\n", WTERMSIG(result));
exit(1);
} else {
printf("Not sure how we exited.\n");
}
return result;
}
int main() {
mySystem("g++ ../random.cpp -o random.out");
mySystem(
("g++ -std=c++11 -o bf.out " + bf).c_str());
mySystem(
("g++ -std=c++11 -o algo.out " + algo).c_str());
for (int T = 1; T <= N; T++) {
// 生成随机数据
mySystem("./random.out > data.in");
// 记录运行的 CPU 时间
double st = clock();
// 暴力解法输出正确答案
mySystem("./bf.out < data.in > data.ans");
double ed = clock();
// 优化解法输出待检查答案
mySystem("./algo.out < data.in > data.out");
// return 0;
if (mySystem("diff data.out data.ans")) {
puts("\033[1;31mWrong Answer\033[0m");
puts("\033[1;31m-----错误数据-------\033[0m");
printf("输入: \n");
mySystem("cat data.in");
return -1;
} else {
printf("\033[1;32mAccepted\033[0m, 测试点 #%d, 用时 %.01f\n", T, ed - st);
}
}
return 0;
}
随机测试样例生成代码(random.cpp):
/*
* @Description:
* @Versions:
* @Author: Vernon Cui
* @Github: https://github.com/vernon97
* @Date: 2021-03-22 20:49:05
* @LastEditors: Vernon Cui
* @LastEditTime: 2021-03-29 19:12:58
* @FilePath: /Projects/random.cpp
*/
#include <iostream>
#include <cstdlib>
#include <string>
#include <sstream>
#include <time.h>
#include "sys/timeb.h"
using namespace std;
int main()
{
struct timeb timeSeed;
ftime(&timeSeed);
srand(timeSeed.time * 1000 + timeSeed.millitm); // milli time
// 产生[a,b)的随机数,可以使用 (rand() % (b-a))+a;
// 产生[a,b]的随机数,可以使用 (rand() % (b-a+1))+a;
// 产生(a,b]的随机数,可以使用 (rand() % (b-a))+a+1;
int n = (rand() % (21 - 1 + 1)) + 1; // 避免前导零
stringstream ss;
ss << (rand() % (10 - 0 + 1)) + 1;
for(int i = 0; i < n; i++)
ss << (rand() % (10 - 0 + 1)) + 0;
cout << ss.str() << endl;
return 0;
}
暴力算法(bf.cpp):
/*
* @Description:
* @Versions:
* @Author: Vernon Cui
* @Github: https://github.com/vernon97
* @Date: 2021-03-22 20:48:38
* @LastEditors: Vernon Cui
* @LastEditTime: 2021-03-29 18:54:47
* @FilePath: /Projects/bf.cpp
*/
#include <iostream>
#include <algorithm>
#include <cstring>
using namespace std;
using ULL = unsigned long long;
const int N = 20;
int n;
int fact[N];
int sumA(ULL A)
{
int sum = 0;
while(A)
{
sum += A % 10;
A /= 10;
}
return sum;
}
ULL findB(ULL A)
{
for(ULL i = A + 1; i < ULLONG_MAX; i++)
if(sumA(i) == sumA(A)) return i;
}
int main()
{
ULL A;
cin >> A;
cout << findB(A) << endl;
return 0;
}
代码(Example.cpp):
#include <iostream>
#include <string>
#include <vector>
#include <algorithm>
#include <sstream>
using namespace std;
vector<int> num;
string tmp;
int main() {
stringstream ss;
cin >> tmp;
// 1. 高精度按位存
for (int i = 0; i < tmp.size(); i++)
{
num.push_back(tmp[i] - '0');
}
if (num.size() == 1)
{
std::cout << num[0] + 9;
return 0;
}
// 2. 倒序查找
int r = num.size() - 1;
int curSum = 0;
while (curSum == 0 || num[r] == 9)
{
if (r == -1) break;
r--;
curSum += num[r + 1];
}
// 3. 处理输出
if (r >= 0) {
for (int i = 0; i < r; i++)
ss << num[i];
ss << num[r] + 1;
curSum--;
int nine_num = static_cast<int>(curSum) / 9;
int num_mod = curSum % 9;
for (int i = r; i < num.size() - 1 - nine_num - 1; i++)
ss << 0;
ss << num_mod;
for (int i = 0; i < nine_num; i++)
ss<< 9;
} else
{
ss << "1";
curSum--;
int nine_num = static_cast<int>(curSum) / 9;
int num_mod = curSum % 9;
r = 0;
for (int i = r; i < static_cast<int>(num.size()) - nine_num - 1; i++)
ss << 0;
ss << num_mod;
for (int i = 0; i < nine_num; i++)
ss << 9;
}
cout << ss.str() << endl;
return 0;
}
测试样例:
以下测试样例为自动生成, 分别为A和B
1023950553073
1023950553082
Accepted, 测试点 #1, 用时 491.0
11410238899918
11410238899927
Accepted, 测试点 #2, 用时 502.0
16099
16189
Accepted, 测试点 #3, 用时 628.0
27617
27626
Accepted, 测试点 #4, 用时 488.0
2021071096708
2021071096717
Accepted, 测试点 #5, 用时 722.0
50893
50929
Accepted, 测试点 #6, 用时 614.0
02 - 编程题02
已知有如下二维离散点:(3.1,1)(4,1.9)(4,4)(6,2.1)(9,7)(9.6,3)(11,2)(12.2,4)(14,2)(18,6),提取其中近似在一条直线上的离散点集群(要求该集群要包含最多的离散点,集群中点到集群对应直线的距离小于0.3)。输出该离散点集群以及对应的直线(以两点表示),给出函数实现,输出结果并说明解题思路(避免使用封装好的库算法)
本题的思路是采用RANSAC来拟合直线;
RANSAC做的一件事就是先随机的选取一对点,用这些点去计算出一个直线,然后用此模型去测试剩余的点,不断重复此过程,直到找到选取的这些数据点集达到了可以接受的程度为止,此时得到的模型便可认为是对数据点的最优模型构建。 答案:
-5.16x + 14.98y + 0.21 = 0
#include <math.h>
#include <fstream>
#include <iostream>
#include <vector>
#include <random>
#include <string>
#include <time.h>
#include "sys/timeb.h"
using namespace std;
const float PI = 3.1415926;
const float MAXD = 0.30;
string filePath = "/Users/vernon/question/question2/sample.in";
int n;
vector<float> x, y;
int max_cnt;
int getNumPoints(float x1, float y1, float x2, float y2) {
float a, b, c, dist;
a = y1 - y2;
b = x2 - x1;
c = y2 * x1 - x2 * y1;
float ab = sqrt(a * a + b * b);
int cnt = 0;
for (int i = 0; i < n; i++) {
dist = abs(a * x[i] + b * y[i] + c) / ab;
if (dist < MAXD) cnt++;
}
return cnt;
}
int main() {
FILE *fin = fopen(filePath.c_str(), "r");
struct timeb timeSeed;
ftime(&timeSeed);
srand(timeSeed.time * 1000 + timeSeed.millitm); // milli time
n = 10;
x.resize(n, 0);
y.resize(n, 0);
// 1. 读取每一个点的坐标
for (int i = 0; i < n; i++)
{
float cur_x, cur_y;
fscanf(fin, "%f %f", &cur_x, &cur_y);
x[i] = cur_x;
y[i] = cur_y;
}
vector<vector<float>> res_pairs;
float x1, y1, x2, y2;
int idx1, idx2, cnt;
float r, theta;
// 2. Ransac 迭代过程
for (int i = 0; i < 100000; i++)
{
idx1 = random() % 10;
do
idx2 = random() % 10;
while (idx2 == idx1);
r = (random() % 300000) / 1000000.0;
theta = (random() % 360000000) / 1000000.0;
theta = theta * PI / 180.;
x1 = x[idx1] + r * cos(theta);
y1 = y[idx1] + r * sin(theta);
r = (random() % 300000) / 1000000.0;
theta = (random() % 361000000) / 1000000.0;
theta = theta * PI / 180.0;
x2 = x[idx2] + r * cos(theta);
y2 = y[idx2] + r * sin(theta);
cnt = getNumPoints(x1, y1, x2, y2);
if (cnt > max_cnt) {
max_cnt = cnt;
res_pairs = {{x1, y1}, {x2, y2}};
}
}
float a, b, c, dist;
x1 = res_pairs[0][0];
y1 = res_pairs[0][1];
x2 = res_pairs[1][0];
y2 = res_pairs[1][1];
a = y1 - y2;
b = x2 - x1;
c = y2 * x1 - x2 * y1;
// 输出最后的直线拟合方程
string line = "直线拟合方程如下: %.2fx";
if (b >= 0)
line += " + %.2fy";
else {
line += " - %.2fy";
b = -b;
}
if (c >= 0)
line += " + %.2f";
else {
line += " - %.2f";
c = -c;
}
line += " = 0\n";
printf(line.c_str(), a, b, c);
return 0;
}
Last updated