The Student Room Group

Set Theory

Hi guys,

I've been asked to show that (RS)1=R1S1(R \cup S)^{-1} = R^{-1} \cup S^{-1}, where RR and SS are binary relations on some set AA.

This is my proof:

{(a,b),(c,d)  (a,b)R1,(c,d)S1}\{ (a, b), (c, d) \ | \ (a, b) \in R^{-1}, (c, d) \in S^{-1} \}
{(a,b),(c,d)  (a,b),(c,d)R1S1}\{ (a, b), (c, d) \ | \ (a, b), (c, d) \in R^{-1} \cup S^{-1} \}
{(a,b),(c,d)  (b,a)R,(d,c)S}\{ (a, b), (c, d) \ | \ (b, a) \in R, (d, c) \in S \}
{(a,b),(c,d)  (b,a),(d,c)RS}\{ (a, b), (c, d) \ | \ (b, a), (d, c) \in R \cup S \}
{(a,b),(c,d)  (a,b),(c,d)(RS)1}\{ (a, b), (c, d) \ | \ (a, b), (c, d) \in (R \cup S)^{-1} \}

Is this a valid proof?
(edited 9 years ago)
Original post by r3l3ntl3ss
Hi guys,

I've been asked to show that (RS)1=R1S1(R \cup S)^-1 = R^-1 \cup S^-1.

This is my proof:

x

Is this a valid proof?

Could you reformat your LaTeX please?
Reply 2
Original post by sarcasmrules
Could you reformat your LaTeX please?


done, had a problem with the curly brackets
Original post by r3l3ntl3ss
done, had a problem with the curly brackets


This is what I would have done:

Multiply both sides by the union of R and S:
(RS)(RS)1=(RS)(R1S1)(R \cup S) \cup (R \cup S)^{-1} = (R \cup S) \cup (R^{-1} \cup S^{-1})

sth operated with sth is e (identity):
e=(RS)(R1S1) e = (R \cup S) \cup (R^{-1} \cup S^{-1})

Union operation is commutative and associative, so:
e=(RR1)(SS1) e = (R \cup R^{-1}) \cup (S \cup S^{-1})

e=ee\therefore e = e \cup e
e=e e = e

I think yours is valid too.
Reply 4
What you've written is false

(RS)cScRc(R \cup S)^c \neq S^c \cup R^c.

Instead (RS)c=ScRc(R \cup S)^c = S^c \cap R^c.

For these kinda of problems where you have a set AA and BB and you want to show that A=BA = B, you have to show inclusion on both sides so that ABA \subseteq B and BA B \subseteq A.

So assume x(RS)cx \in (R \cup S)^c and prove that xScRcx \in S^c \cap R^c also and vice versa.
(edited 9 years ago)
Reply 5
Original post by sarcasmrules
This is what I would have done:

Multiply both sides by the union of R and S:
(RS)(RS)1=(RS)(R1S1)(R \cup S) \cup (R \cup S)^{-1} = (R \cup S) \cup (R^{-1} \cup S^{-1})

sth operated with sth is e (identity):
e=(RS)(R1S1) e = (R \cup S) \cup (R^{-1} \cup S^{-1})

Union operation is commutative and associative, so:
e=(RR1)(SS1) e = (R \cup R^{-1}) \cup (S \cup S^{-1})

e=ee\therefore e = e \cup e
e=e e = e

I think yours is valid too.

(RS)(RS)c(R \cup S) \cup (R \cup S)^c is not ee. In fact AAc=U A \cup A^c = U, where UU can be taken as the universe where AUA \subseteq U.

Also ee as an identity makes no sense in set theory. There's no axiom of set theory that forces you to have an identity element in a set.
(edited 9 years ago)
Original post by 0x2a
(RS)(RS)c(R \cup S) \cup (R \cup S)^c is not ee. In fact AAc=U A \cup A^c = U, where UU can be taken as the universe where AUA \subseteq U.

Also ee as an identity makes no sense in set theory. There's no axiom of set theory that forces you to have an identity element in a set.

Oh crap, I read the title as "group theory" :colondollar:

I see where you're coming from. AcAA^c \cup A would be any value in the Universe.
(edited 9 years ago)
Reply 7
Ah I also forgot to mention that RR and SS are binary relations on some set AA.

Is it still the same principle, though?
Reply 8
Original post by r3l3ntl3ss
Ah I also forgot to mention that RR and SS are binary relations on some set AA.

Is it still the same principle, though?


The thing is now that you say it's a binary relation (which I assume is symmetric), and defined such that R1:={(b,a)(a,b)R}R^{-1} := \{(b,a)|(a,b) \in R\}, then (RS)1=R1S1(R \cup S)^{-1} = R^{-1} \cup S^{-1} is correct.

Consider (b,a)(RS)1((a,b)R)((a,b)S)(b,a) \in (R \cup S)^{-1} \Rightarrow ((a,b) \in R) \vee ((a,b) \in S) \Rightarrow ((b,a)R1)((b,a)S1)(b,a)R1S1((b,a) \in R^{-1}) \vee ((b,a) \in S^{-1}) \Rightarrow (b,a) \in R^{-1} \cup S^{-1}, and so (RS)1R1S1(R \cup S)^{-1} \subseteq R^{-1} \cup S^{-1}. And you can show the converse showing that (RS)1=R1S1(R \cup S)^{-1} = R^{-1} \cup S^{-1}.
(edited 9 years ago)

Quick Reply

Latest