Take the 2-minute tour ×
Computer Science Stack Exchange is a question and answer site for students, researchers and practitioners of computer science. It's 100% free, no registration required.

I'm reviewing for finals and have a sample problem that I think I understand, but would like someone to bless my understanding or smack me and tell me why I'm wrong.

I'm presented with a problem $\Pi$ of unknown complexity class. If I can transform $\Pi$ to some problem $X$, where $X \in {\sf P}$, what does that tell me about $\Pi$?

I think allows me to conclude that $\Pi \in {\sf P}$, right? If I can reduce $\Pi$ to another problem that's deterministically solvable in polynomial time, and the transformation itself can be done "easily" in polynomial time, then I can conclude that $\Pi$ is deterministically solvable in polynomial time, and therefore that $\Pi \in {\sf P}$ correct?

Conversely, given the same input, transforming $X$ to $\Pi$ in polynomial time allows me to conclude nothing meaningful, since nothing is known about $\Pi$ right?

share|improve this question
add comment

2 Answers 2

up vote 3 down vote accepted

In essence, you are right. Given a problem $\Pi$, if you can find a poly-time reduction from $\Pi$ to any problem $x\in P$, then you have proof that $\Pi\in P$ as well.

A reduction from problem $A$ to $B$ is basically a function that transforms any input of $A$ into an input of $B$, such that the solutions agree. That means, you can use any algorithm for $B$ to solve $A$.

For the second part, if you can reduce a problem $x\in P$ to $\Pi$, then $\Pi$ might be in $P$, or not. However, if you can find a way to reduce all problems in $P$ (or equivalently, some $P$-complete problem), then you have shown that $\Pi$ is $P$-hard.

share|improve this answer
add comment

Right, if you have an algorithm A that reduces $\Pi$ to $x$ in polytime, and you can solve $x$ is polytime using some algorithm B, then you can solve $\Pi$ in polytime by applying first A and then B.

Using polytime reductions on a problem in P however is rather boring, since you can solve the instance during the reduction and map every instance of $x$ to either a trivial yes or a trivial no answer. Normally you restrict yourself to logspace reductions for problems from P.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.