Proof of points arbitrarily close to supremum

Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
1 replies · 2K views
Vale132
Messages
7
Reaction score
0

Homework Statement



Let [itex]S \subset \mathbb{R}[/itex] be bounded above. Prove that [itex]s \in \mathbb{R}[/itex] is the supremum of [itex]S[/itex] iff. [itex]s[/itex] is an upper bound of [itex]S[/itex] and for all [itex]\epsilon > 0[/itex], there exists [itex]x \in S[/itex] such that [itex]|s - x| < \epsilon[/itex].

Homework Equations


**Assume I have only the basic proof methods, some properties of inequalities, and the Completeness Axiom at my disposal**

Assume I have already proved that if [itex]s \in \mathbb{R}[/itex] is the supremum of [itex]S[/itex], then for all [itex]\epsilon > 0[/itex], there exists [itex]x \in S[/itex] such that [itex]s - x < \epsilon[/itex].

The Attempt at a Solution


The [itex]\Rightarrow[/itex] direction:
By the definition of the supremum, [itex]s[/itex] is an upper bound.

Thus [itex]s \geq x[/itex] for all [itex]x \in S[/itex], so [itex]s - x \geq 0[/itex]. Then [itex]|s - x| = s - x < \epsilon[/itex], by the result mentioned above.

The [itex]\Leftarrow[/itex] direction:
Suppose that [itex]s \in \mathbb{R}[/itex] is an upper bound of [itex]S[/itex] and that for all [itex]\epsilon > 0[/itex], there exists [itex]x \in S[/itex] such that [itex]|s - x| < \epsilon[/itex]. Now suppose for a contradiction that [itex]s[/itex] is not the supremum of [itex]S[/itex]. Then there exists an upper bound [itex]u[/itex] of [itex]S[/itex] with [itex]u < s[/itex].

But since [itex]S[/itex] is bounded above, it does have a supremum. Call the supremum [itex]t[/itex]. We have that [itex]t \leq s[/itex], and for every [itex]\epsilon > 0[/itex], there exists [itex]y \in S[/itex] such that [itex]|t - y| < \epsilon[/itex].I'm wondering whether there are any logical errors so far, and whether someone could give me a tiny hint as to whether I'm on the right track, and what I might think about next. I'm posting here rather than Math.SE because I would prefer "Socratic" help rather than an answer. Thanks!
 
on Phys.org
Vale132 said:

Homework Statement



Let [itex]S \subset \mathbb{R}[/itex] be bounded above. Prove that [itex]s \in \mathbb{R}[/itex] is the supremum of [itex]S[/itex] iff. [itex]s[/itex] is an upper bound of [itex]S[/itex] and for all [itex]\epsilon > 0[/itex], there exists [itex]x \in S[/itex] such that [itex]|s - x| < \epsilon[/itex].

Homework Equations


**Assume I have only the basic proof methods, some properties of inequalities, and the Completeness Axiom at my disposal**

Assume I have already proved that if [itex]s \in \mathbb{R}[/itex] is the supremum of [itex]S[/itex], then for all [itex]\epsilon > 0[/itex], there exists [itex]x \in S[/itex] such that [itex]s - x < \epsilon[/itex].

The Attempt at a Solution


The [itex]\Rightarrow[/itex] direction:
By the definition of the supremum, [itex]s[/itex] is an upper bound.

Thus [itex]s \geq x[/itex] for all [itex]x \in S[/itex], so [itex]s - x \geq 0[/itex]. Then [itex]|s - x| = s - x < \epsilon[/itex], by the result mentioned above.

The [itex]\Leftarrow[/itex] direction:
Suppose that [itex]s \in \mathbb{R}[/itex] is an upper bound of [itex]S[/itex] and that for all [itex]\epsilon > 0[/itex], there exists [itex]x \in S[/itex] such that [itex]|s - x| < \epsilon[/itex]. Now suppose for a contradiction that [itex]s[/itex] is not the supremum of [itex]S[/itex]. Then there exists an upper bound [itex]u[/itex] of [itex]S[/itex] with [itex]u < s[/itex].

But since [itex]S[/itex] is bounded above, it does have a supremum. Call the supremum [itex]t[/itex]. We have that [itex]t \leq s[/itex], and for every [itex]\epsilon > 0[/itex], there exists [itex]y \in S[/itex] such that [itex]|t - y| < \epsilon[/itex].I'm wondering whether there are any logical errors so far, and whether someone could give me a tiny hint as to whether I'm on the right track, and what I might think about next. I'm posting here rather than Math.SE because I would prefer "Socratic" help rather than an answer. Thanks!
The ⇒ direction seems correct.

For the ⇐ direction:
You have ##s## and the supremum ##t##.
If ##s=t##, you are done.
##s<t## is impossible, as ##t## is the supremum, the least upper bound.
That leaves the case ##s>t##.
Now try to prove that then there must exist an ##x \in S## satisfying ##x>t##, which contradicts ##t## being a supremum.
Hint: set ##\epsilon=\frac{s-t}{3}##.
 
Last edited: