The approximability of different NP-hard problems

48 Views Asked by At

I'm fairly new to the topic Computational Complexity and had the following question (I therefore apologies before hand for any poorly stated terminology).

Suppose i have two optimization problems $A$ and $B$ and suppose further that $A$ is polynomial time reducible to $B$ i.e $A \leq_{p} B$. Therefore i know if $A$ is NP-hard that $B$ is also NP-hard.

If however i have more information about $A$ for example

1.$A$ is APX-hard does that also mean that $B$ is APX-hard?

2.$B$ admits a PTAS does that also mean that $A$ admits a PTAS or can it be APX-hard?

Thanks

1

There are 1 best solutions below

0
On

In general APX-hardness of $A$ will not immediately imply APX-hardness of $B$. Nevertheless, the implication holds if you strengthen the conditions of the reduction. If the reduction is approximation reserving, i.e. an approximation solution to $B$ corresponds to an approximation solution to $A$ when the reduction is applied, the implication still holds. The wikipedia article on approximation-preserving reductions might give you a more formal idea of this concept.