T2真是一个折磨人的小妖精,写了两版DFS,第二版计数DFS才过。T3是三分模板,感觉也可以求导数。T4的数据规模才n=1000,因此中心扩展的 O ( n 2 ) O(n^2) O(n2)当仁不让。
滑窗经典题
从某个左端点出发,按顺序找到最远的右端点
然后把该右端点变成新的左端点,继续寻找直至结束
import java.io.*;
import java.util.*;
public class Main {
public static void main(String[] args) {
Scanner sc = new Scanner(new BufferedInputStream(System.in));
int n = sc.nextInt();
int[] arr = new int[n];
for (int i = 0; i < n; i++) arr[i] = sc.nextInt();
int ans = 0;
int j = 0;
while (j < n) {
int k = j + 1;
while (k < n && Math.abs(arr[k] - arr[k - 1]) <= 1) {
k++;
}
ans = Math.max(ans, k - j);
j = k;
}
System.out.println(ans);
}
}
#include
using namespace std;
int main() {
int n;
cin >> n;
vector arr(n);
for (int i = 0; i< n; i++) {
cin >> arr[i];
}
int res = 1;
int j = 0;
while (j < n) {
int k = j + 1;
while (k < n && abs(arr[k - 1] - arr[k]) <= 1) {
k++;
}
res = max(res, k - j);
j = k;
}
cout << res << endl;
return 0;
}
暴力搜索好像TLE了,而且需要额外处理去重
那如何搜索,可以不考虑去重的情况呢?
可以对每个字母进行频数统计
然后进行dfs,这样有一个好处,就是天然去重,时间复杂度为 O ( n ! ) O(n!) O(n!)
import java.io.*;
import java.util.*;
public class Main {
// 计数DFS
public static int dfs(int[] nums, int n, int now, int prev) {
if (now == n) {
return 1;
}
int res = 0;
for (int i = 0; i < nums.length; i++) {
if (nums[i] > 0 && i != prev) {
nums[i]--;
res += dfs(nums, n, now + 1, i);
nums[i]++;
}
}
return res;
}
public static void main(String[] args) {
Scanner sc = new Scanner(new BufferedInputStream(System.in));
char[] str = sc.next().toCharArray();
Map<Character, Integer> hash = new HashMap<>();
for (char c: str) hash.merge(c, 1, Integer::sum);
int[] nums = hash.values().stream().mapToInt(x -> x).toArray();
System.out.println(dfs(nums, str.length, 0, -1));
}
}
#include
using namespace std;
using int64 = long long;
int64 dfs(const string &s, char last, int mask) {
int n = s.length();
if (mask == (1 << n) - 1) {
return 1LL;
}
int64 res = 0;
for (int i = 0; i < n; i++) {
if ((mask & (1 << i)) == 0) {
if (s[i] == last || (i > 0 && s[i] == s[i - 1] && (mask &(1 << (i - 1))) == 0) ) {
continue;
}
res += dfs(s, s[i], mask | (1 << i));
}
}
return res;
}
int main() {
string s;
cin >> s;
sort(s.begin(), s.end());
int64 res = dfs(s, ' ', 0);
cout << res << endl;
return 0;
}
三分板子题,该函数为凹函数。
z = y / (v + x*t) + t
该函数先单调递减,然后单调递增
import java.io.BufferedInputStream;
import java.util.Scanner;
public class Main {
static double calc(double y, double x, double v, double t) {
return y / (v + x * t) + t;
}
public static void main(String[] args) {
Scanner sc = new Scanner(new BufferedInputStream(System.in));
double v = sc.nextDouble(), x = sc.nextDouble(), y = sc.nextDouble();
// 三分板子
double l = 0.0, r = 1e9;
while (r - l >= 1e-7) {
// 要求1e-6
double s = (r - l) / 3.0;
double t1 = l + s;
double t2 = l + 2 * s;
double res1 = calc(y, x, v, t1);
double res2 = calc(y, x, v, t2);
if (res1 >= res2) {
l = t1;
} else {
r = t2;
}
}
System.out.println(calc(y, x, v, l));
}
}
f(t) = y / (v + x*t) + t
求导
f‘(t) = -xy / (v + x*t)^2 + 1
求f’(t) = 0 的解
x^2 * t^2 + 2vx*t + v^2 - xy = 0; t为变量, x,y,v都是常数
t’=(-v +/- sqrt(xy)) / x,
不知道,牛客啥时候对latex语法不那么友好了,写起来难受,看的人更难受
因为xy>0, 所以一定有解,同时 t>=0, 因此只有一个可能解
t’ = (-v + sqrt(xy)) / x
t’ = max(0, t’), 保证t>=0
import java.io.BufferedInputStream;
import java.util.Scanner;
public class Main {
static double calc(double y, double x, double v, double t) {
return y / (v + x * t) + t;
}
public static void main(String[] args) {
Scanner sc = new Scanner(new BufferedInputStream(System.in));
double v = sc.nextDouble(), x = sc.nextDouble(), y = sc.nextDouble();
// 求一元二次方程的根
double t = (-v + Math.sqrt(x*y)) / x;
// 保证t>=0
t = Math.max(0, t);
System.out.println(calc(y, x, v, t));
}
}
因为n=1000,所以中心扩展就好了
import java.io.BufferedInputStream;
import java.util.Scanner;
public class Main {
public static void main(String[] args) {
Scanner sc = new Scanner(new BufferedInputStream(System.in));
int n = sc.nextInt();
long[] arr = new long[n];
for (int i = 0; i < arr.length; i++) {
arr[i] = sc.nextInt();
}
long ans = 0;
long mod = 10_0000_0007l;
for (int i = 0; i < n; i++) {
// 枚举某个区域
long t = arr[i] * (arr[i] + 1) / 2;
ans = (ans + t % mod) % mod;
// 中心扩展
for (int j = 1; i - j >= 0 && i + j < n; j++) {
if (arr[i - j] != arr[i + j]) {
// 长度不等,取最小,通过跳出
ans = (ans + Math.min(arr[i - j], arr[i + j])) % mod;
break;
} else {
ans = (ans + arr[i - j]) % mod;
}
}
}
System.out.println(ans);
}
}