1
$\begingroup$

I need to prove that if $P=NP$ then every two non trivial languages (meaning not $\Sigma^*$ or $\varnothing$) $A,B\in coNP$ holds $A \equiv_{P} B$

(this means that $A \leq_{P} B$ and also $B \leq_{P} A$)

Heres what I've done so far:

assume $P=NP$ and let $A,B\in coNP$, so $\overline{A}, \overline{B} \in NP$,

since $P=NP$ we get that $\overline{A}, \overline{B} \in P$, and since P is closed under complement we get that $A,B\in P$

since $P \subseteq NP$ we get that $A,B\in NP$

Now I don't know how to polynomialy reduce one to the other. I thought about using a NP-complete language like $3-SAT$ but the question is too general

$\endgroup$

1 Answer 1

2
$\begingroup$

We know that if $P=NP$, then we essentially have $P = NP = coNP$ (as you have already demonstrated). Now the problem boils down to showing that any two problems in $P$ are poly-time reducible to each other.

To show that $A \le_p B$ for any $A, B \in P$, we have to essentially exhibit the existence of a mapping function $f$ that converts any input $I_A$ of $A$ into an input $I_B$ of $B$ in poly-time. Here's how we can design such a fuction $f$:

Pick two instances $I_B^{(true)}$ and $I_B^{(false)}$ of $B$ such that one of them is a true instance and the other one is a false instance. This can be determined in polytime since $B \in P$. Now take the input $I_A$ and actually solve it in poly-time (recall $A \in P$). If the result is true, return $I_B = I_B^{(true)}$; otherwise, return $I_B = I_B^{(false)}$.

Also, you may see these related discussions: Problem about $NP \neq coNP$, Can any problem in P be converted to any other problem in P in polynomial time?.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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