11095 - Tabriz City

[joeluchoa]

Is it possible to solve this problem with graph coloring(two colors)???
I've used this and got WA!!

[smile2ka10]

no it is not bicoloring!
It is vertex cover problem.

[Ferdous]

I think it would have been a bicoloring problem if the constraint was "If a junction has a tourist information centre, no adjacent junction can have another one" instead of "at least one of two junctions at the end of each street have a tourist information center"

[arsalan_mousavian]

try this input

Code: Select all

1
15
20
0 1
1 2
2 3
3 4
4 5
5 6
6 7
7 3
6 3
6 4
8 9
9 10 
10 11
11 12
11 13
12 13
13 14
8 14
8 11
9 14

the answer should be:

Code: Select all

Case #1: 8
1 3 5 6 9 11 13 14

hope it helps

[joeluchoa]

Ok, AC now!!
Thanks

[fernando]

After i have read what vertex cover problem is, i agree this problem asks for the vertex cover, but how can you compute it?, specially when this is a NP problem whit n=30, do you have to make some kind of backtracking with pruning or something like that?, thanks in advance

[mamun]

Probably I haven't understood the problem clearly. Aren't we supposed to find the minimum number of information centers needed to form so that all the junctions have a center either at it or adjacent to it?

If so, for the graph posted by arsalan_mousavian 6 centers should be enough, I think. Don't 1 3 5 8 10 12 take care of all the junctions?

[Hisoka]

He wants to do this in a way that at least one of two junctions at the end of each street have a tourist information center

[mamun]

So it's the streets that must ensure that at least one end has a center. Yes, my previous assumption would have missed streets (4 6) and some more. Thank you Hisoka.

[Vexorian]

It is NP complete but the input size is quite small. BTW, if you read the graph as an array of edges the solution is kind of straight forward

[nymo]

What kind of optimization do you do in this problem? I have represented the graph using adjacency matrix; I am using bit field for marking... I am getting TLE, the code is basically a bruteforce search ... I select a node, mark all edges incident to it and then try other nodes, after returing from other nodes I simpy unmark the edges and try starting with other nodes... Can you share some of your tricks??? Thanks.

[MIB]

thanks to Vexorian... I got accepted easily by reading the graph as an array of edges...

the solution is obviously brute force search...
nymo, The only optimization that I did, was: whenever I decided not to take a vertex as an information centre, I marked all the adjacent vertices as taken. Hope it will help...

[nymo]

Thanks to MIB and Vexorian, I am now getting WAs... I try some test cases... they are okay. Can someone post some more test cases???
Thanks.

[nymo]

Thanks MIB. I 've got ACC

[okz]

help me plz...
i use the approach of computing the nodes combination from small to large
in another word,try taking all possible 1 node and test ,then try taking 2 nodes combination,then 3...and more
but it become very slow after 8 nodes combination
i just got lots of TLE in online-judge...
plz tell me how should i design the brute force approach
will it be better if i use binary searching?
thx.