常用的解题技巧:尺取法
尺取法:顾名思义,像尺子一样取一段,借用挑战书上面的话说,尺取法通常是对数组保存一对下标,即所选取的区间的左右端点,然后根据实际情况不断地推进区间左右端点以得出答案。之所以需要掌握这个技巧,是因为尺取法比直接暴力枚举区间效率高很多,尤其是数据量大的
时候,所以尺取法是一种高效的枚举区间的方法,一般用于求取有一定限制的区间个数或最短的区间等等。当然任何技巧都存在其不足的地方,有些情况下尺取法不可行,无法得出正确答案。
使用尺取法时应清楚以下四点:
1、 什么情况下能使用尺取法? 2、何时推进区间的端点? 3、如何推进区间的端点? 3、何时结束区间的枚举?
尺取法通常适用于选取区间有一定规律,或者说所选取的区间有一定的变化趋势的情况,通俗地说,在对所选取区间进行判断之后,我们可以明确如何进一步有方向地推进区间端点以求解满足条件的区间,如果已经判断了目前所选取的区间,但却无法确定所要求解的区间如何进一步
得到根据其端点得到,那么尺取法便是不可行的。首先,明确题目所需要求解的量之后,区间左右端点一般从最整个数组的起点开始,之后判断区间是否符合条件在根据实际情况变化区间的端点求解答案。
以下是几个经典的使用尺取法的例题,都是从挑战书上引用的。(尺取法通常会需要对某些量进行预处理,以便能在使用时快速地判断。)
(代码via博主consciousman)
1、 Poj3061
题意:给定一个序列,找出最短的子序列长度,使得其和大于或等于S。
分析:首先,序列都是正数,如果一个区间其和大于等于S了,那么不需要在向后推进右端点了,因为其和也肯定大于等于S但长度更长,所以,当区间和小于S时右端点向右移动,和大于等于S时,左端点向右移动以进一步找到最短的区间,如果右端点移动到区间末尾其和还不大于等于S,
结束区间的枚举。
这个题目区间和明显是有趋势的:单调变化,所以根据题目要求很容易求解,但是在使用之间需要对区间前缀和进行预处理计算。
代码:
- #include <cstdio>
- #include <algorithm>
- #include <cstring>
- #define MAX 100005
- #define LL long long
- #define INF 0x3f3f3f3f
- using namespace std;
- LL a[100010];
- int n, t, ans = INF;
- LL sum, s;
- int main()
- {
- scanf("%d", &t);
- while (t--){
- scanf("%d %I64d", &n, &s);
- for (int i = 0; i < n; i++) scanf("%I64d", a+i);
- int st = 0, en = 0;
- ans = INF; sum = 0;
- while (1){
- while (en<n && sum<s) sum += a[en++];
- if (sum < s) break;
- ans = min(ans, en-st);
- sum -= a[st++];
- }
- if (ans == INF) ans = 0;
- printf("%d\n", ans);
- }
- return 0;
- }
2、 poj3320
题意:一本书有P页,每一页都一个知识点,求去最少的连续页数覆盖所有的知识点。
分析:和上面的题一样的思路,如果一个区间的子区间满足条件,那么在区间推进到该处时,右端点会固定,左端点会向右移动到其子区间,且其子区间会是更短的,只是需要存储所选取的区间的知识点的数量,那么使用map进行映射以快速判断是否所选取的页数是否覆盖了所有的
知识点。
代码:
- #include <cstdio>
- #include <algorithm>
- #include <cstring>
- #include <set>
- #include <map>
- #define MAX 1000010
- #define LL long long
- #define INF 0x3f3f3f3f
- using namespace std;
- int a[MAX];
- map <int, int> cnt;
- set <int> t;
- int p, ans = INF, st, en, sum;
- int main()
- {
- scanf("%d", &p);
- for (int i = 0; i < p; i++) scanf("%d", a+i), t.insert(a[i]);
- int num = t.size();
- while (1){
- while (en<p && sum<num)
- if (cnt[a[en++]]++ == 0) sum++;
- if (sum < num) break;
- ans = min(ans, en-st);
- if (--cnt[a[st++]] == 0) sum--;
- }
- printf("%d\n", ans);
- return 0;
- }
这个代码虽然是直接copy的,但仔细研读发现也是很有料的。
首先是要理解题意,你要保证涵盖每个知识点,那你先要知道由多少个知识点吧?而这里用set来存储然后求set的size()即可。
那么就要考虑如何尺取了,应该看涵盖知识点的个数sum吧?那问题来了,可能会存在知识点重复的情况,这里的解决方法是用map,在end后移或者start后移的时候其实都是去判断这个知识点的map值是否为0 。这里的处理还是很有意思的!
3、 poj2566
题意:给定一个数组和一个值t,求一个子区间使得其和的绝对值与t的差值最小,如果存在多个,任意解都可行。
分析:明显,借用第一题的思路,既然要找到一个子区间使得和最接近t的话,那么不断地找比当前区间的和更大的区间,如果区间和已经大于等于t了,那么不需要在去找更大的区间了,因为其和与t的差值更大,然后区间左端点向右移动推进即可。所以,首先根据计算出所有的区间和,
排序之后按照上面的思路求解即可。
代码:
- #include <cstdio>
- #include <algorithm>
- #include <cstring>
- #define INF 0x3f3f3f3f
- #define LL long long
- #define MAX 100010
- using namespace std;
- typedef pair<LL, int> p;
- LL a[MAX], t, ans, tmp, b;
- int n, k, l, u, st, en;
- p sum[MAX];
- LL myabs(LL x)
- {
- return x>=0? x:-x;
- }
- int main()
- {
- while (scanf("%d %d", &n, &k), n+k){
- sum[0] = p(0, 0);
- for (int i = 1; i <= n; i++){
- scanf("%I64d", a+i);
- sum[i] = p(sum[i-1].first+a[i], i);
- }
- sort(sum, sum+1+n); //sort函数支持线性序列,支持pair数组的。(不支持map的)
- while (k--){
- scanf("%I64d", &t);
- tmp = INF; st = 0, en = 1;
- while(en <= n){
- b = sum[en].first-sum[st].first;
- if(myabs(t-b) < tmp){
- tmp = myabs(t-b);
- ans = b;
- l = sum[st].second; u = sum[en].second;
- }
- if(b > t) st++;
- else if(b < t) en++;
- else break;
- if(st == en) en++;
- }
- if (u < l) swap(u, l);
- printf("%I64d %d %d\n", ans, l+1, u);
- }
- }
- return 0;
- }
这道题也是有一定的技术含量的,用到了“前缀和”。
其实在分析这道问题的时候就会发现:很可能一个>0的区间和之后加上个负数,那么是不是可能和t的差值减小了,那么我end的动作就很难定(因为对这题来说,如果区间和b<t,start和end该怎样变化能去寻找会不会有更小的abs(t-b)? end向后移那万一之后是个负数,b反而还减小了,abs(t-b)还增大了;start向后移那万一删掉了一个正数,那么b也是反而减少了)————》》所有的矛盾都可以由一个条件解决,那就是做“前缀和”,然后从小到大排序,通过前缀和取差值来得区间和b。这样的话,我能保证我目前的这个区间和b,在end后移之后b是铁定增大的,start后移之后b是铁定减小的。(可以联想递增的柱状图)
因为题目要求的是满足条件的最小区间,肯定需要min函数取一下,记录最佳的区间的端点。要注意!:因为用了前缀和之后排序了,所以标号的位置是打乱了的,这里就很巧妙的用到了“元组pair”来用first保存前缀和,second保存当前端点index。由于前缀和排序后端点值也被打乱了,不过前缀和都呈包含关系,最后把两个端点由小到大输出即可。
4、 poj2739&poj2100
题意:找到某一个区间使得区间内的数的和/平方和等于某一给定值k。
分析:很明显了,几乎之与上面的poj2566又是一样的,当区间右端点不能再向右推进且区间和仍小于k的话就可以结束区间的枚举了。
代码:
poj2739;
- #include <cstdio>
- #include <cstring>
- #include <algorithm>
- #include <vector>
- #include <utility>
- #include <queue>
- #define INF 0x3f3f3f3f
- #define LL long long
- using namespace std;
- int prime[] = {2,3,5,7,11,13,17,19,23,29,31,37,41,43,47,53,59,61,67,71,73,79,83,89,97,101,103,107,109,113,127,131,137,139,149,151,157,163,167,173,179,181,191,193,197,199,211,223,227,229,233,239,241,251,257,263,269,271,277,281,283,293,307,311,313,317,331,337,347,349,353,359,367,373,379,383,389,397,401,409,419,421,431,433,439,443,449,457,461,463,467,479,487,491,499,503,509,521,523,541,547,557,563,569,571,577,587,593,599,601,607,613,617,619,631,641,643,647,653,659,661,673,677,683,691,701,709,719,727,733,739,743,751,757,761,769,773,787,797,809,811,821,823,827,829,839,853,857,859,863,877,881,883,887,907,911,919,929,937,941,947,953,967,971,977,983,991,997,1009,1013,1019,1021,1031,1033,1039,1049,1051,1061,1063,1069,1087,1091,1093,1097,1103,1109,1117,1123,1129,1151,1153,1163,1171,1181,1187,1193,1201,1213,1217,1223,1229,1231,1237,1249,1259,1277,1279,1283,1289,1291,1297,1301,1303,1307,1319,1321,1327,1361,1367,1373,1381,1399,1409,1423,1427,1429,1433,1439,1447,1451,1453,1459,1471,1481,1483,1487,1489,1493,1499,1511,1523,1531,1543,1549,1553,1559,1567,1571,1579,1583,1597,1601,1607,1609,1613,1619,1621,1627,1637,1657,1663,1667,1669,1693,1697,1699,1709,1721,1723,1733,1741,1747,1753,1759,1777,1783,1787,1789,1801,1811,1823,1831,1847,1861,1867,1871,1873,1877,1879,1889,1901,1907,1913,1931,1933,1949,1951,1973,1979,1987,1993,1997,1999,2003,2011,2017,2027,2029,2039,2053,2063,2069,2081,2083,2087,2089,2099,2111,2113,2129,2131,2137,2141,2143,2153,2161,2179,2203,2207,2213,2221,2237,2239,2243,2251,2267,2269,2273,2281,2287,2293,2297,2309,2311,2333,2339,2341,2347,2351,2357,2371,2377,2381,2383,2389,2393,2399,2411,2417,2423,2437,2441,2447,2459,2467,2473,2477,2503,2521,2531,2539,2543,2549,2551,2557,2579,2591,2593,2609,2617,2621,2633,2647,2657,2659,2663,2671,2677,2683,2687,2689,2693,2699,2707,2711,2713,2719,2729,2731,2741,2749,2753,2767,2777,2789,2791,2797,2801,2803,2819,2833,2837,2843,2851,2857,2861,2879,2887,2897,2903,2909,2917,2927,2939,2953,2957,2963,2969,2971,2999,3001,3011,3019,3023,3037,3041,3049,3061,3067,3079,3083,3089,3109,3119,3121,3137,3163,3167,3169,3181,3187,3191,3203,3209,3217,3221,3229,3251,3253,3257,3259,3271,3299,3301,3307,3313,3319,3323,3329,3331,3343,3347,3359,3361,3371,3373,3389,3391,3407,3413,3433,3449,3457,3461,3463,3467,3469,3491,3499,3511,3517,3527,3529,3533,3539,3541,3547,3557,3559,3571,3581,3583,3593,3607,3613,3617,3623,3631,3637,3643,3659,3671,3673,3677,3691,3697,3701,3709,3719,3727,3733,3739,3761,3767,3769,3779,3793,3797,3803,3821,3823,3833,3847,3851,3853,3863,3877,3881,3889,3907,3911,3917,3919,3923,3929,3931,3943,3947,3967,3989,4001,4003,4007,4013,4019,4021,4027,4049,4051,4057,4073,4079,4091,4093,4099,4111,4127,4129,4133,4139,4153,4157,4159,4177,4201,4211,4217,4219,4229,4231,4241,4243,4253,4259,4261,4271,4273,4283,4289,4297,4327,4337,4339,4349,4357,4363,4373,4391,4397,4409,4421,4423,4441,4447,4451,4457,4463,4481,4483,4493,4507,4513,4517,4519,4523,4547,4549,4561,4567,4583,4591,4597,4603,4621,4637,4639,4643,4649,4651,4657,4663,4673,4679,4691,4703,4721,4723,4729,4733,4751,4759,4783,4787,4789,4793,4799,4801,4813,4817,4831,4861,4871,4877,4889,4903,4909,4919,4931,4933,4937,4943,4951,4957,4967,4969,4973,4987,4993,4999,5003,5009,5011,5021,5023,5039,5051,5059,5077,5081,5087,5099,5101,5107,5113,5119,5147,5153,5167,5171,5179,5189,5197,5209,5227,5231,5233,5237,5261,5273,5279,5281,5297,5303,5309,5323,5333,5347,5351,5381,5387,5393,5399,5407,5413,5417,5419,5431,5437,5441,5443,5449,5471,5477,5479,5483,5501,5503,5507,5519,5521,5527,5531,5557,5563,5569,5573,5581,5591,5623,5639,5641,5647,5651,5653,5657,5659,5669,5683,5689,5693,5701,5711,5717,5737,5741,5743,5749,5779,5783,5791,5801,5807,5813,5821,5827,5839,5843,5849,5851,5857,5861,5867,5869,5879,5881,5897,5903,5923,5927,5939,5953,5981,5987,6007,6011,6029,6037,6043,6047,6053,6067,6073,6079,6089,6091,6101,6113,6121,6131,6133,6143,6151,6163,6173,6197,6199,6203,6211,6217,6221,6229,6247,6257,6263,6269,6271,6277,6287,6299,6301,6311,6317,6323,6329,6337,6343,6353,6359,6361,6367,6373,6379,6389,6397,6421,6427,6449,6451,6469,6473,6481,6491,6521,6529,6547,6551,6553,6563,6569,6571,6577,6581,6599,6607,6619,6637,6653,6659,6661,6673,6679,6689,6691,6701,6703,6709,6719,6733,6737,6761,6763,6779,6781,6791,6793,6803,6823,6827,6829,6833,6841,6857,6863,6869,6871,6883,6899,6907,6911,6917,6947,6949,6959,6961,6967,6971,6977,6983,6991,6997,7001,7013,7019,7027,7039,7043,7057,7069,7079,7103,7109,7121,7127,7129,7151,7159,7177,7187,7193,7207,7211,7213,7219,7229,7237,7243,7247,7253,7283,7297,7307,7309,7321,7331,7333,7349,7351,7369,7393,7411,7417,7433,7451,7457,7459,7477,7481,7487,7489,7499,7507,7517,7523,7529,7537,7541,7547,7549,7559,7561,7573,7577,7583,7589,7591,7603,7607,7621,7639,7643,7649,7669,7673,7681,7687,7691,7699,7703,7717,7723,7727,7741,7753,7757,7759,7789,7793,7817,7823,7829,7841,7853,7867,7873,7877,7879,7883,7901,7907,7919,7927,7933,7937,7949,7951,7963,7993,8009,8011,8017,8039,8053,8059,8069,8081,8087,8089,8093,8101,8111,8117,8123,8147,8161,8167,8171,8179,8191,8209,8219,8221,8231,8233,8237,8243,8263,8269,8273,8287,8291,8293,8297,8311,8317,8329,8353,8363,8369,8377,8387,8389,8419,8423,8429,8431,8443,8447,8461,8467,8501,8513,8521,8527,8537,8539,8543,8563,8573,8581,8597,8599,8609,8623,8627,8629,8641,8647,8663,8669,8677,8681,8689,8693,8699,8707,8713,8719,8731,8737,8741,8747,8753,8761,8779,8783,8803,8807,8819,8821,8831,8837,8839,8849,8861,8863,8867,8887,8893,8923,8929,8933,8941,8951,8963,8969,8971,8999,9001,9007,9011,9013,9029,9041,9043,9049,9059,9067,9091,9103,9109,9127,9133,9137,9151,9157,9161,9173,9181,9187,9199,9203,9209,9221,9227,9239,9241,9257,9277,9281,9283,9293,9311,9319,9323,9337,9341,9343,9349,9371,9377,9391,9397,9403,9413,9419,9421,9431,9433,9437,9439,9461,9463,9467,9473,9479,9491,9497,9511,9521,9533,9539,9547,9551,9587,9601,9613,9619,9623,9629,9631,9643,9649,9661,9677,9679,9689,9697,9719,9721,9733,9739,9743,9749,9767,9769,9781,9787,9791,9803,9811,9817,9829,9833,9839,9851,9857,9859,9871,9883,9887,9901,9907,9923,9929,9931,9941,9949,9967,9973};
- int main()
- {
- int n;
- while (scanf("%d", &n), n){
- int ans, st, en, sum;
- st = en = ans = sum = 0;
- while (1){
- if (sum == n) ans++;
- if (sum >= n) sum -= prime[st++];
- else{
- if (prime[en] <= n) sum += prime[en++];
- else break;
- }
- }
- printf("%d\n", ans);
- }
- }
poj2100:
- #include <cstdio>
- #include <cstring>
- #include <algorithm>
- #include <vector>
- #include <utility>
- #include <queue>
- #define INF 0x3f3f3f3f
- #define LL long long
- using namespace std;
- typedef pair<LL, pair<LL, LL> > p;
- p ans[1010];
- int main()
- {
- LL n, st, en, sum;
- while (~scanf("%I64d", &n)){
- st = 1, en = 1, sum = 0;
- int k = 0;
- while (1){
- if (sum == n) ans[k++] = p(en-st, pair<LL, LL>(st, en-1));
- if (sum >= n) sum -= st*st, st++;
- else{
- if (en*en <= n) sum += en*en, en++;
- else break;
- }
- }
- printf("%d\n", k);
- for (int i = 0; i < k; i++){
- printf("%I64d ", ans[i].first);
- for (int j = ans[i].second.first; j <= ans[i].second.second; j++) printf("%I64d ", j);
- puts("");
- }
- }
- return 0;
- }
其实就我的理解来说,尺取法有一个关键的思考角度是:
start和end能否通过后移可以让当前情况朝着更佳的方向发展?如果不能,是不是该进行些数据处理?
比如在全是正数的数组中,求区间使得区间和b=t,那么b<t了肯定end后移可以让b增大靠近t,b>=t了肯定start后移可以让b减小寻找下一个区间。
对于那道求最短区间覆盖所有知识点的,肯定不能仅仅start和end后移可以让当前情况朝着更佳的方向发展,所以考虑到加入map的处理来辅助计数知识点个数。
对于那道求区间和使b和t的差值最小的,当b<t或b>t时,不管start和end后移,都不能确定说这个操作能够使b更靠近t,所以想到用“前缀和作差”的方法解决了这个问题。
(2018/4/15)补充一道算法导论上的求“最大子数组”的题:
第一种方法,需要遍历,时间复杂度为o(n^2):
#include<iostream>
#include<bits/stdc++.h>
using namespace std;
int main()
{
int A[6];
for(int i=0;i<6;i++)
{
cin>>A[i];
}
for(int i=1;i<6;i++)
{
A[i]=A[i-1]+A[i]; //前缀和
}
int maxx=-1e6;
int l=0,r=0; //记录最大子数组的起点终点
for(int i=0;i<6;i++)
{
int ans=A[i]; //最大子数组要么是A[0...i]
int ll=0,rr=i;
for(int j=0;j<i;j++)
{
if(ans<A[i]-A[j]) //要么是A[j+1...i]
{
ans=A[i]-A[j];
ll=j+1;
rr=i;
}
}
if(maxx<ans)
{
maxx=ans;
l=ll;
r=rr;
}
}
cout<<"最大子数组的区间是"<<l<<" "<<r<<",和是"<<maxx<<endl;
return 0;
}
第二种方法,都不算尺取法,只算一种前缀和的方法,把前缀和算出来排个序,然后针对3种情况进行分析(注意,最大区间和不是简单地最大前缀和-最小前缀和(之前想的那个梯形图,那个是对于绝对值的),比如序列“-1 -1 -1”,最大区间和并不是-1-(-3)=2!),所以要分三种情况讨论:
#include<iostream>
#include<bits/stdc++.h>
using namespace std;
//前缀和+尺取法
int main()
{
int A[6];
for(int i=0;i<6;i++)
{
cin>>A[i];
}
pair<int,int> a[6]; //first表示前缀和,second表示当前下标
a[0].first=A[0];
a[0].second=0;
for(int i=1;i<6;i++)
{
a[i].first=a[i-1].first+A[i];
a[i].second=i;
}
int l=0,r=0; //记录最大子数组的起点终点
sort(a,a+6); //给前缀和从小到大排序!
int maxx;
if(a[0].first>0) //若前缀和都为正数,则取最大的前缀和
{
maxx=a[5].first;
l=0;
r=a[5].second;
}
else if(a[5].first<=0)//若前缀和都<=0,则取最大的前缀和
{
maxx=a[5].first;
l=0;
r=a[5].second;
}
else //若a[0].first<0而a[5].first>0
{
if(a[5].second<a[0].second) //若最大的前缀和为小标,如“1 -2 -3”序列
{
maxx=a[5].first;
l=0;
r=a[5].second;
}
else //若最大的前缀和为大标,那么从最大前缀和-最小前缀和为最大的区间和,如“-5 6 7”序列
{
maxx=a[5].first-a[0].first;
l=a[0].second+1;
r=a[5].second;
}
}
cout<<"最大子数组的区间是"<<l<<" "<<r<<",和是"<<maxx<<endl;
return 0;
}