10740 - Not the Best

[david]

What reference do you need? Isn't slightly modified Dijkstra (with counting the k-th shortest path to all the verticles) good enough?

No, it isn't. It is not inmediately clear how to deal with zero-length cycles.

[Krzysztof Duleba]

I don't really get what's wrong with zero-length circuits. You should visit a verticle only if it's k-th shortest path from the source isn't already known. So if you have zero-length circuit, you should take it at most k times.

[Monsoon]

Hi,
could anyone give more test cases to check my program...,
is there exist any special test case?

TIA.

[Monsoon]

thx, i found my mistake...

[chuzpa]

I removed my code :$, It has been long enogh here haha, enjoy the problem ..

[chuzpa]

Hi nobody answered me ... :s hahaha but it is accepted now ....

[ImLazy]

I use the MPS algorithm introduced in the article The K shortest paths problem from this page. It's fast.