通过将数拆分成一位位的进行dp(记忆化搜索),状态最基本的有:位置($pos$),最高位限制($limit$),前导零($lead$),前一位($pre$)等等,通常需要的状态视题目而定。
记忆化搜索的数组$dp$由多维构成,每一位都是一种状态的因素。
数位$dp$入门题,$dp$数组为$dp[pos][pre][limit]$($limit$可以通过更改代码去掉这一维)。
#include<iostream> #include<cstdio> #include<cstring> #include<string> #include<algorithm> #include<cstdlib> #include<cmath> using namespace std; const int N = 10; int as[N], bs[N], lena, lenb; long long dp[N][N][2]; int a, b; inline int read(){ int i = 0, f = 1; char ch = getchar(); for(; (ch < ‘0‘ || ch > ‘9‘) && ch != ‘-‘; ch = getchar()); if(ch == ‘-‘) f = -1, ch = getchar(); for(; ch >= ‘0‘ && ch <= ‘9‘; ch = getchar()) i = (i << 3) + (i << 1) + (ch - ‘0‘); return i * f; } inline void wr(long long x){ if(x < 0)putchar(‘-‘), x = -x; if(x > 9) wr(x / 10); putchar(x%10 + ‘0‘); } inline long long dfs(int pos, int last, bool limit, int *s, int l){ if(pos == l + 1) return 1; if(dp[pos][last][limit] != -1) return dp[pos][last][limit]; int high = limit ? s[pos] : 9; long long ret = 0; for(int i = 0; i <= high; i++){ if(last == 6 && i == 2) continue; if(i == 4) continue; ret += dfs(pos + 1, i, limit && (i == high), s, l); } dp[pos][last][limit] = ret; return ret; } int main(){ while(1){ a = read(), b = read(); if(!a && !b) break; //---------------------------------- memset(dp, -1, sizeof dp); memset(as, 0, sizeof as); lena = 0; if(a > b) swap(a, b); int tmp; tmp = a - 1; while(tmp){ as[++lena] = tmp % 10; tmp /= 10; } reverse(as + 1, as + lena + 1); long long r1 = dfs(1, 0, 1, as, lena); //---------------------------------- memset(dp, -1, sizeof dp); memset(bs, 0, sizeof bs); lenb = 0; tmp = b; while(tmp){ bs[++lenb] = tmp % 10; tmp /= 10; } reverse(bs + 1, bs + lenb + 1); long long r2 = dfs(1, 0, 1, bs, lenb); //------------------------------------- wr(r2 - r1), putchar(‘\n‘); } }
同样是入门级别的题,$dp$中加入前导零这一维(亦可更改代码去掉),当当前位可作为前导零时,有两种选择:
若不能在作为前导$0$,则必须满足差值大于等于$2$的要求。
#include<iostream> #include<cstdio> #include<cstdlib> #include<cstring> #include<string> #include<algorithm> #include<vector> using namespace std; const int N = 15; int dp[N][N][2][2]; //位置,前一个数,最高位限制 int as[N], bs[N], lena, lenb, a, b; inline int dfs(int pos, int last, bool limit, bool lead, int *s, int l){ if(pos == l + 1) return 1; if(dp[pos][last][limit][lead] != -1) return dp[pos][last][limit][lead]; int high = limit ? s[pos] : 9, ret = 0; for(int i = 0; i <= high; i++){ if(!lead){ if(i != 0)ret += dfs(pos + 1, i, limit && i == high, 1, s, l); else ret += dfs(pos + 1, i, limit && i == high, 0, s, l); } else if(abs(i - last) >= 2) ret += dfs(pos + 1, i, limit && i == high, 1, s, l); } dp[pos][last][limit][lead] = ret; return ret; } int main(){ scanf("%d%d", &a, &b); if(a > b) swap(a, b); //------------------------ memset(dp, -1, sizeof dp); int tmp = a - 1; while(tmp) as[++lena] = tmp % 10, tmp /= 10; reverse(as + 1, as + lena + 1); int r1 = dfs(1, 0, 1, 0, as, lena); memset(dp, -1, sizeof dp); tmp = b; while(tmp) bs[++lenb] = tmp % 10, tmp /= 10; reverse(bs + 1, bs + lenb + 1); int r2 = dfs(1, 0, 1, 0, bs, lenb); cout<<r2 - r1<<endl; return 0; }
学到了。本题难点在于离散压空间。若正常的进行$dp$,那么数组大小早就爆了。
现在考虑这样一个问题:对于$1~9$的数而言,最小公倍数为2520. 由题知,设最后的数为$x$
数学推到知$a \equiv b (mod p) \Leftrightarrow q | a - b$
原文:http://www.cnblogs.com/CzYoL/p/7231573.html