10405 - Longest Common Subsequence

[emotional blind]

Can you please make a hint for me to feel the difference between bottom up and top down in this problem. o(n^2) solution of this kind:

Code: Select all

for i in sequence 1
  for j in sequence 2
    if s1[i]=s2[j] then
      ans[i][j] = ans[i-1][j-1] + 1; 
    else if .... and so on ...

is top-down? or bottom-up?
thanx in advance

BOTTOM UP

[andysoft]

oh, shame on me
then... it is the faster one, right?
would you mind if I write you some stuff on e-mail? it is too confidential for the forum

[emotional blind]

oh, shame on me
then... it is the faster one, right?
would you mind if I write you some stuff on e-mail? it is too confidential for the forum

okay I dont mind.. email me at any time..
use this subject: 10405

[mccarre]

I'm not doing it with DP yet. I'm almost positive my method works, and want to see it fail, before resorting to taking the DP approach.
I've tried mixing the strings so that the largest subsequences can be mixed in a random order, and i've tried tripling that approach, and I've tried adding random throw off characters, but I can't break my program. However it gets WA.

Here are 2 inputs I've used, first being the shortest complex string I could think of, and the second being the same with a lot of random characters in between:

Code: Select all

abcdxzyacdxzyb
acdxzybabcdxzy
a436432b5c346d23x6z56y54a62c6dx64665426z2y6b26
uaiopciudopxipzioypoibpuaobpcioodpiuxpozpy

Here's my code:

Code: Select all

#include <iostream>
#include <string>
#include <vector>

using namespace std;		//status: completed, again they giving fucking terrible input, so obviously I can't get it accepted.

string s1, s2;				//flip it around and compare it again? Tyr this and evaluate!
int maxsub; 	//also wiki longest common subsequence.
int s2iter; int s1iter;

void countinstuff()
{
	int max=1000000000, dist=0,temp,s2loc; bool foundone;

	/*find next occurance of a letter*/
	int i;
	for(i=0; i<s1.length(); i++)
		if(s2iter=s2.find(s1[i])+1) //this means it doesn't find anything... better then ==string::npos
			break;
	
	//cout << "Found first match at " << i<<", "<<(s2iter-1)<<endl;
	/***/

	for(int k=i; k<s1.length(); k++)
	{
			int bleh=k;
			foundone=false;
			max=1000000000;
			//cout << "k now equals " << k << endl;
			for(int j=k; j<s1.length(); j++)
			{
				/*find next occurance of a letter*/
				for(i=j; i<s1.length(); i++)
					if(temp=s2.find(s1[i],s2iter)+1) //this means it doesn't find anything... better then ==string::npos
						break;
				//cout << "Found a match at " << i<<", "<<(temp-1)<<endl;
				j=i;
				if(j==s1.length() && !foundone)	k=s1.length();
				else			foundone=true;
				/***/
				//cerr << "max = " << max << " & last dist = " << dist << endl;
				if(j!=s1.length())
				{	if((dist=(j-bleh)+(temp-s2iter))<max)
					{
						//cerr << "found better match at " <<j<<", "<<temp-1<<"while respectivly, k and s2iter are: "<<k<<", "<<s2iter<<endl;
						max=dist;
						s2loc=temp;
						k=j;
					}}
			}
			s2iter=s2loc;
			//cout << "Found best match at " << k<<", "<<(s2iter-1)<<endl;
			if(foundone)	maxsub++;
			//cout << "maxsub now equals " << maxsub << endl;
	}
	//cout << "**************************"<<endl;
}

int main()
{
while(cin >> s1)
{
	cin >> s2;

	s2iter=0;
	maxsub=1;
	//cout << "s1=" << s1 << " s2=" << s2<<endl;
	countinstuff();
	//cout << "maxsub=" << maxsub << " (before swap)"<<endl;

	cout << maxsub << endl;
}
return 0;
}

All i need is a complex input to break my code. Preferably not too long, so it isn't too hard to trace ^^.

[serur]

Please, use the existing thread.
As for the problem, you should read strings with fgets() (there can be blanks)

[vkubushyn]

Hi I also get WA, and have no idea why. My program handles spaces and blank lines correctly and I think my LCS algorithm does not have a flaw. Please help.

Code: Select all

#include <iostream>
#include <string>
#include <stdio.h>
using namespace std;

char s1[1005];
char s2[1005];
void findLCS();
int main() {
	while(gets(s1) && gets(s2))
		findLCS();
	return 0;
}

void findLCS() {
	int m = strlen(s1)-1;
	int n = strlen(s2)-1;
	int C[m+1][n+1];
	for(int i = 0; i <= m; i++) 
		C[i][0] = 0;
	for(int i = 1; i <= n; i++)
		C[0][i] = 0;
	for(int i = 0; i < m; i++) {
		for(int j = 0; j < n; j++) {
			if(s1[i] == s2[j])
				C[i+1][j+1] = C[i][j]+1;
			else
				C[i+1][j+1] = (C[i+1][j] > C[i][j+1]) ? C[i+1][j] : C[i][j+1];
		}
	}
	cout << C[m][n] << endl;
}

Thanks in advance.

[rio]

Try this:

Code: Select all

a
a

-----
Rio

[vkubushyn]

Hmmm, I get 1 for that input. That's correct isn't it?

[rio]

In my platform and in judge's platform, you code outputs 0 with the input.
The newline is "\n", not "\r\n".

-----
Rio

[vkubushyn]

Wow, ok it got AC now. Thanks a lot, I'll know not to count the '\r' anymore. I hate it when a correct program gets WA because of a compatibility error ;(.

[Obaida]

Why I got WA?

Code: Select all

Problem Solved

[apurba]

maximum string length is 1000.........check it

[Obaida]

Thank you i got accepted before.
Best of luck.

[ashish_thakur]

Why is it giving run time error please help?

#include <string.h>
#include <stdio.h>


int
LCSlength(char *seq1, char *seq2){

int len1 = strlen(seq1);
int len2 = strlen(seq2);
if(!(len1 && len2))
return 0;

int C[len1][len2];
int i=0,j=0;
int cond=0;
for(;i<len2;i++)
if(!cond)
cond=C[0]=(seq1[0]==seq2);
else
C[0]=cond;

cond=0;
for(;j<len1;j++)
if(!cond)
cond=C[j][0]=(seq1[j]==seq2[0]);
else
C[j][0]=cond;

for(i=1;i<len1;i++)
for(j=1;j<len2;j++)
if (seq1==seq2[j])
C[j] = C[i-1][j-1] + 1;
else
C[j] = C[j-1]>C[i-1][j]?C[j-1]:C[i-1][j];

return C[len1-1][len2-1];
}


int main(){

char seq1[1005];
char seq2[1005];
while(strlen(gets(seq1))&& strlen(gets(seq2)))
printf("%d\n",LCSlength(seq1,seq2));
return 0;
}

[mf]

gets() returns NULL on end-of-file.
strlen expects a non-NULL pointer, and so it crashes when gets() passes NULL to it.