• Aucun résultat trouvé

Characterization of the supremum

N/A
N/A
Protected

Academic year: 2022

Partager "Characterization of the supremum"

Copied!
5
0
0

Texte intégral

(1)

MAT246H1-S – LEC0201/9201

Concepts in Abstract Mathematics

REAL NUMBERS – 2

March 16th, 2021

Jean-Baptiste Campesato MAT246H1-S – LEC0201/9201 – Mar 16, 2021 1 / 5

(2)

Characterization of the supremum

Proposition

Let𝐴 ⊂ ℝand𝑀 ∈ ℝ. Then𝑀 =sup(𝐴) ⇔ {

∀𝑥 ∈ 𝐴, 𝑥 ≤ 𝑀

∀𝜀 > 0, ∃𝑥 ∈ 𝐴, 𝑀 − 𝜀 < 𝑥

𝐴 ℝ

𝑀 𝑀 − 𝜀

𝜀

𝑥

Proof.

Assume that𝑀 =sup(𝐴). Then𝑀is an upper bound of𝐴so∀𝑥 ∈ 𝐴, 𝑥 ≤ 𝑆.

We know that if𝑇is an other upper bound of𝐴then𝑀 ≤ 𝑇(since𝑀is the least upper bound).

So, by taking the contrapositive, if𝑇 < 𝑀then𝑇isn’t an upper bound of𝐴.

Let𝜀 > 0. Since𝑀 − 𝜀 < 𝑀, we know that𝑀 − 𝜀is not an upper bound of𝐴, meaning that there exists𝑥 ∈ 𝐴such that𝑀 − 𝜀 < 𝑥.

Assume that {

∀𝑥 ∈ 𝐴, 𝑥 ≤ 𝑀

∀𝜀 > 0, ∃𝑥 ∈ 𝐴, 𝑀 − 𝜀 < 𝑥

Then, by the first condition,𝑀is an upper bound of𝐴. Let’s prove it is the least one.

We will show the contrapositive: if𝑇 < 𝑀then𝑇isn’t an upper bound of𝐴.

Let𝑇 ∈ ℝ. Assume that𝑇 < 𝑀. Set𝜀 = 𝑀 − 𝑇 > 0. Then there exists𝑥 ∈ 𝐴such that𝑀 − 𝜀 < 𝑥, i.e.𝑇 < 𝑥.

Hence𝑇isn’t an upper bound of𝐴.

Jean-Baptiste Campesato MAT246H1-S – LEC0201/9201 – Mar 16, 2021 2 / 5

(3)

Characterization of the infimum

Proposition

Let𝐴 ⊂ ℝand𝑚 ∈ ℝ. Then

𝑚 =inf(𝐴) ⇔ {

∀𝑥 ∈ 𝐴, 𝑚 ≤ 𝑥

∀𝜀 > 0, ∃𝑥 ∈ 𝐴, 𝑥 < 𝑚 + 𝜀

𝐴 ℝ

𝑚 𝑚 + 𝜀

𝜀

𝑥

Jean-Baptiste Campesato MAT246H1-S – LEC0201/9201 – Mar 16, 2021 3 / 5

(4)

ℝ is archimedean

Theorem: ℝ is archimedean

∀𝜀 > 0, ∀𝐴 > 0, ∃𝑛 ∈ ℕ, 𝑛𝜀 > 𝐴 Proof.

Let𝜀 > 0and𝐴 > 0.

Assume by contradiction that∀𝑛 ∈ ℕ, 𝑛𝜀 ≤ 𝐴.

Then𝐸 = {𝑛𝜀 ∶ 𝑛 ∈ ℕ}is non-empty and bounded from above so it admits a supremum 𝑀 =sup𝐸by the least upper bound principle.

Since𝑀 − 𝜀 < 𝑀,𝑀 − 𝜀is not an upper bound of𝐸, so there exists𝑛 ∈ ℕsuch that𝑛𝜀 > 𝑀 − 𝜀.

Therefore(𝑛 + 1)𝜀 > 𝑀, hence a contradiction. ■

The above theorem means that lim

𝑛→+∞

1

𝑛 = 0, or equivalently thatℝdoesn’t contain infinitesimal elements (i.e. there is not infinitely large or infinitely small elements).

Jean-Baptiste Campesato MAT246H1-S – LEC0201/9201 – Mar 16, 2021 4 / 5

(5)

Floor function

Proposition: floor function

For every𝑥 ∈ ℝ, there exists a unique𝑛 ∈ ℤsuch that𝑛 ≤ 𝑥 < 𝑛 + 1.

We say that𝑛is the integer part (or the floor function value) of𝑥and we denote it by⌊𝑥⌋.

Proof.Let𝑥 ∈ ℝ.

Existence.

First case: if𝑥 ≥ 0.

We set𝐸 = {𝑛 ∈ ℕ ∶ 𝑥 < 𝑛}.

By the archimedean property (with𝜀 = 1), there exists𝑚 ∈ ℕsuch that𝑚 > 𝑥. Hence𝐸 ≠ ∅.

By the well-ordering principle,𝐸admits a least element𝑝.

We have that𝑥 < 𝑝since𝑝 ∈ 𝐸and that𝑝 − 1 ≤ 𝑥since𝑝 − 1 ∉ 𝐸.

Therefore𝑛 = 𝑝 − 1satisfies𝑛 ≤ 𝑥 < 𝑛 + 1.

Second case: if𝑥 < 0.Then we apply the first case to−𝑥.

Uniqueness.Assume that𝑛, 𝑛∈ ℤare two suitable integers. Then(1) 𝑛 ≤ 𝑥 < 𝑛 + 1and𝑛≤ 𝑥 < 𝑛+ 1.

We deduce from the last inequality that(2) − 𝑛− 1 < −𝑥 ≤ −𝑛. Summing(1)and(2), we get that𝑛 − 𝑛− 1 < 0 < 𝑛 − 𝑛+ 1.

Hence𝑛 − 𝑛< 1, i.e.𝑛 − 𝑛≤ 0, and−1 < 𝑛 − 𝑛i.e.0 ≤ 𝑛 − 𝑛.

Therefore𝑛 = 𝑛. ■

Jean-Baptiste Campesato MAT246H1-S – LEC0201/9201 – Mar 16, 2021 5 / 5

Références

Documents relatifs

This last result is actually a particular case of [9], where it is proved that for any non creeping Lévy process, the law of the first passage time over x &gt; 0 is always

Conclude to the existence and uniqueness of a global (in time) variational solution to the nonlinear McKean- Vlasov equation..

In particular, in Lemma 4.1 we prove the limit (1.1) when A is the number of marked nodes, and we deduce the convergence of a critical GW tree conditioned on the number of marked

Heuristic arguments suggest that the original version of Theorem 1.2 is true, but at this time we do not have a rigorous argument to back up this claim.. We would like to thank

the distribution of the supremum for compound recurrent processes and for stochastic processes with stationary independent increments.. The paper also contains some

In any 3-dimensional warped product space, the only possible embedded strictly convex surface with constant scalar curvature is the slice sphere provided that the embedded

We construct a cohomology theory on a category of finite digraphs (directed graphs), which is based on the universal calculus on the algebra of functions on the vertices of

Here we present a local discontinuous Galerkin method for computing the approximate solution to the kinetic model and to get high order approximations in time, space and