poj_1743 后缀数组

题目大意

    给定一串数字,长度为N。定义数字中的某个连续的子串为一个"theme",只要子串满足: 
(1)长度 >= 5 
(2)和该子串相同或者该子串的“变种串”在整串数字中出现次数大于1 
(3)假设整串中有k个该子串及其“变种串”,那么其中至少有两个不相重叠 
    求满足要求的 "theme" 串的最长长度。

题目分析

    (1)首先考虑将“变种”串和原子串相互比较的问题,对字符串中所有索引大于等于1的字符都用该字符减去前一个字符,这样得到串的差串之后,原theme和其“变种”就一样了,此时只需要求差串中的最长相同子串,且这些子串之间不重叠

    求最长相同子串,可以考虑使用后缀数组和height数组。显然,height越大,则两个子串的公共前缀越长,越有可能是最长相同子串。但是,题目对"theme"串的要求(3)至少两个不重叠,因此需要考虑height[i]在尽可能大的同时,保证SA[i]和SA[i-1]之间的差值要大于height[i]以保证不重叠

    (2)然后,试图求解是否存在长度为M的"theme"串。 
容易看出,后缀Suffix(j)和Suffix(k)的最长公共前缀的长度为 height[rank[j]+1], height[rank[j]+2]...height[rank[k]]的最小值。i从1到N遍历,通过height[i]>=M将i分开,即将后缀分成若干组,每组中的后缀的公共前缀长度均大于等于M,且可以肯定组A中的某后缀t1和组B中的某后缀t2的公共前缀长度小于M。若存在这样的组,则可以确定找到了公共前缀大于等于M的子串,下一步需要确定这些子串不重叠。只需要在组内寻找 SA[i] 之间最大的查看,看是否大于子串的长度,若大于则可以确定不重叠。

    (3)最后,求解"theme"串长度M的最大值,用二分法对"theme"串的可能长度进行二分求解,长度范围为0到N。每次二分得到中值M,先判断能否找到长度为k的"theme"串,若不能,则减小M,否则增加M。直到找到长度M最大的"theme"串。

实现(c++)

#define _CRT_SECURE_NO_WARNINGS
#include<stdio.h>
#include<string.h>
#define MAX(a, b) a>b? a:b
#define MAX_ARRAY_SIZE 20005
#define LETTERS 10000

int gStrLen;
int gStr[MAX_ARRAY_SIZE];
int gCount[MAX_ARRAY_SIZE];
int gSuffixArray[MAX_ARRAY_SIZE];
int gRank[MAX_ARRAY_SIZE];
int gOrderBySecondKey[MAX_ARRAY_SIZE];
int gFirstKeyArray[MAX_ARRAY_SIZE];
int gHeight[MAX_ARRAY_SIZE];

bool Compare(int* arr, int a, int b, int step){
	return arr[a] == arr[b] && arr[a + step] == arr[b + step];
}


void GetStr(char* str){
	memset(gStr, 0, sizeof(gStr));
	gStrLen = strlen(str);
	for (int i = 0; i < gStrLen; i++){
		gStr[i] = str[i] - 'a' + 1;
	}
	gStr[gStrLen++] = 0;
}

void GetSuffixArray(){
	int n = gStrLen;
	memset(gCount, 0, sizeof(gCount));
	for (int i = 0; i < n; i++){
		gRank[i] = gStr[i];
		gCount[gRank[i]] ++;
	}
	for (int i = 1; i < LETTERS; i++){
		gCount[i] += gCount[i - 1];
	}
	for (int i = n - 1; i >= 0; i--){
		gSuffixArray[--gCount[gRank[i]]] = i;
	}
	int step = 1;
	int* rank = gRank, *order_by_second_key = gOrderBySecondKey;
	int m = LETTERS;
	while (step < n){
		int p = 0;
		for (int i = n - step; i < n; i++){
			order_by_second_key[p++] = i;
		}
		for (int i = 0; i < n; i++){
			if (gSuffixArray[i] >= step){
				order_by_second_key[p++] = gSuffixArray[i] - step;
			}
		}
		for (int i = 0; i < n; i++){
			gFirstKeyArray[i] = rank[order_by_second_key[i]];
		}
		for (int i = 0; i < m; i++){
			gCount[i] = 0;
		}
		for (int i = 0; i < n; i++){
			gCount[gFirstKeyArray[i]] ++;
		}
		for (int i = 1; i < m; i++){
			gCount[i] += gCount[i - 1];
		}
		for (int i = n - 1; i >= 0; i--){
			gSuffixArray[--gCount[gFirstKeyArray[i]]] = order_by_second_key[i];
		}
		int* tmp = rank;
		rank = order_by_second_key;
		order_by_second_key = tmp;
		rank[gSuffixArray[0]] = 0;
		p = 0;
		for (int i = 1; i < n; i++){
			if (Compare(order_by_second_key, gSuffixArray[i], gSuffixArray[i - 1], step)){
				rank[gSuffixArray[i]] = p;
			}
			else{
				rank[gSuffixArray[i]] = ++p;
			}
		}
		m = p + 1;
		step *= 2;
	}
}

void GetHeight(){
	int n = gStrLen;
	for (int i = 1; i < n; i++){
		gRank[gSuffixArray[i]] = i;
	}
	int k = 0, j;
	gHeight[0] = 0;
	for (int i = 0; i < n - 1; i++){
		j = gSuffixArray[gRank[i] - 1];
		if (k){
			k--;
		}
		while (i + k < n && j + k < n && gStr[i + k] == gStr[j + k]){
			k++;
		}
		gHeight[gRank[i]] = k;
	}
}
bool Find(int k){
	int end = 1;
	int min_pos, max_pos;
	while (end < gStrLen){
		max_pos = min_pos = gSuffixArray[end-1];
		while (end < gStrLen && gHeight[end] >= k - 1){
			if (min_pos > gSuffixArray[end]){
				min_pos = gSuffixArray[end];
			}
			if (max_pos < gSuffixArray[end]){
				max_pos = gSuffixArray[end];
			}

			end ++;
		}
		if (max_pos - min_pos >= k){
			return true;
		}
		end ++;
	}
	return false;
}
void printstr(int n){
	printf("string = 
");
	for (int i = 0; i < n; i++){
		printf("%d ", gStr[i]);
	}
	printf("
");
}
void printsuffix(int n){
	printf("suffix = 
");
	for (int i = 0; i < n; i++){
		printf("%d ", gSuffixArray[i]);
	}
	printf("
");
}
void printheigt(int n){
	printf("height = 
");
	for (int i = 0; i < n; i++){
		printf("%d ", gHeight[i]);
	}
	printf("
");
}
int main(){
	int n;
	while (true){
		scanf("%d", &n);
		
		if (n == 0){
			break;
		}

		for (int i = 0; i < n; i++){
			scanf("%d", &gStr[i]);
		}
		int min = 100;
		for (int i = 1; i < n; i++){
			gStr[i - 1] = gStr[i] - gStr[i - 1];
			min = gStr[i - 1] < min ? gStr[i - 1] : min;
		}
		min--;
		for (int i = 0; i < n; i++){
			gStr[i] -= min;
		}
		gStr[n-1] = 0;
		gStrLen = n;

		GetSuffixArray();
		GetHeight();
//		printstr(n);
//		printsuffix(n);
//		printheigt(n);
		int beg = 0, end = n, mid, max;
		bool flag = true;
		while (beg < end){
			mid = (beg + end) / 2;
			if (Find(mid)){
				beg = mid + 1;
				max = mid;
			}
			else{
				if (mid <= 5){
					flag = false;
					break;
				}
				end = mid;
			}
		}
		if (!flag){
			printf("0
");
		}
		else{
			printf("%d
", max);
		}
	}
	return 0;
}
原文地址:https://www.cnblogs.com/gtarcoder/p/4835535.html