We defined $\mathbb{Q}$ as the set of equivalence classes for the relation $\sim$. Tentatively define operations $+,\cdot:\mathbb{Q}\times\mathbb{Q}\rightarrow\mathbb{Q}$ by
$[(a,b)]+[(c,d)] = [(ad+bc,bd)],$ $[(a,b)]\cdot[(c,d)]=[(ac,bd)]$
These definitions have a potential problem. They read: "Given equicalence classes $x$" and $y$, choose elements $(a,b)$ and $(c,d)$ of them, perform such and such operations, and let $x+y$ or $x\cdot y$ be the equivalence class of the result." But what if one chose different representatives $(a',b')$ and $(c',d')$? Might those choices produce a different class?
Prove that the definitions are indeed sound by showing that, for all pairs in $A$,
$(a,b)\sim(a',b')$ and $(c,d)\sim(c',d') \implies \begin{cases} (ad+bc,bd)\sim(a'd'+b'c',b'd')\\ (ac,bd)\sim(a'c',b'd') \end{cases}$
I am given that the equivalence relation is $(m,n)\sim (p,q)$ if $mq=np$ in the set $A=\{(m,n) \in \mathbb{Z}\times\mathbb{Z}:n\ne0 \}$.
I am not sure how using representatives $(a',b')$ and $(c',d')$ would product
Remember, $(a, b) \sim (a', b')$ means that $ab' = a'b$.
You want to show that
$(a,b)\sim(a',b') \text{ and } (c,d)\sim(c',d') \implies \begin{cases} (ad+bc,bd)\sim(a'd'+b'c',b'd')\\ (ac,bd)\sim(a'c',b'd') \end{cases} $
I'll do the second one.
We want to show that, if $(a,b)\sim(a',b') \text{ and } (c,d)\sim(c',d') $ then $(ac,bd)\sim(a'c',b'd') $.
From the definition of "$\sim$", this means that $acb'd' = a'c'bd $. Since $(a,b)\sim(a',b') $, $ab' = a'b$. Similarly, $cd' = c'd$. Multiplying these, $ab'cd' = a'bc'd$, which is what you want, rearranged.
The first one, for addition, is more complicated, so I'll leave it for you.