Why is showing that ILP in NP not trivial

129 Views Asked by At

I have a question regarding the topic of showing that ILP is in NP

What is the problem with Guess and Check?

Guess a solution and then check if it is optimal. Or further: Calculate a solution via Simplex and then check if it is optimal. Why is that not sufficient for showing ILP in NP?

Thank you!

1

There are 1 best solutions below

0
On BEST ANSWER

NP refers to decision problems rather than optimization problems. A version of ILP that is in NP is

Given integer matrix $A \in \mathbb Z^{m \times n}$ and vector $b \in \mathbb Z^m$ does there exist a vector $x \in \mathbb Z^n$ such that $A x \le b$?

It is in NP because we can verify a "yes" answer in polynomial time: guess such an $x$ and verify that $Ax \le b$.

On the other hand, there does not seem to be any way to verify a "no" answer in polynomial time, i.e. this version of ILP is not believed to be in co-NP. Thus the following version of ILP is not believed to be in NP:

Given integer matrix $A \in \mathbb Z^{m \times n}$, vector $b \in \mathbb Z^m$, vector $c \in \mathbb Z^n$ and integer $d$, is it true that $\max \{c^T x: x \in \mathbb Z^n, A x \le b\} = d$?