Loading [MathJax]/jax/element/mml/optable/BasicLatin.js
Skip to main content
Library homepage
 

Text Color

Text Size

 

Margin Size

 

Font Type

Enable Dyslexic Font
Mathematics LibreTexts

3.3: Intervals in Eⁿ

( \newcommand{\kernel}{\mathrm{null}\,}\)

Consider the rectangle in E2 shown in Figure 2. Its interior (without the perimeter consists of all points (x,y)E2 such that

a1<x<b1 and a2<y<b2;

i.e.,

x(a1,b1) and y(a2,b2).

Thus it is the Cartesian product of two line intervals, (a1,b1) and (a2,b2). To include also all or some sides, we would have to replace open intervals by closed, half-closed, or half-open ones. Similarly, Cartesian products of three line intervals yield rectangular parallelepipeds in E3. We call such sets in En intervals.

Screen Shot 2019-05-29 at 11.31.34 PM.png

Definition

1. By an interval in En we mean the Cartesian product of any n intervals in E1 (some may be open, some closed or half-open, etc.).

2. In particular, given

¯a=(a1,,an) and ¯b=(b1,,bn)

with

akbk,k=1,2,,n,

we define the open interval (¯a,¯b), the closed interval [¯a,¯b], the half-open interval (¯a,¯b], and the half-closed interval [¯a,¯b) as follows:

(¯a,¯b)={¯x|ak<xk<bk,k=1,2,,n}=(a1,b1)×(a2,b2)××(an,bn)[¯a,¯b]={¯x|akxkbk,k=1,2,,n}=[a1,b1]×[a2,b2]××[an,bn](¯a,¯b]={¯x|ak<xkbk,k=1,2,,n}=(a1,b1]×(a2,b2]××(an,bn][a,b)={¯x|akxk<bk,k=1,2,,n}=[a1,b1)×[a2,b2)××[an,bn)

In all cases, ¯a and ¯b are called the endpoints of the interval. Their distance

ρ(¯a,¯b)=|¯b¯a|

is called its diagonal. The n differences

bkak=k(k=1,,n)

are called its n edge-lengths. Their product

nk=1k=nk=1(bkak)

is called the volume of the interval (in E2 it is its area, in E1 its length) .\) The point

¯c=12(¯a+¯b)

is called its center or midpoint. The set difference

[¯a,¯b](¯a,¯b)

is called the boundary of any interval with endpoints ¯a and b; it consists of 2n "faces" defined in a natural manner. (How?)

We often denote intervals by single letters, e.g.. A=(¯a,¯b), and write dA for "diagonal of A and vA or vol A for "volume of A . " If all edge-lengths b_{k}-a_{k} are equal, A is called a cube (in E^{2}, a square). The interval A is said to be degenerate iff b_{k}=a_{k} for some k, in which case, clearly,

\operatorname{vol} A=\prod_{k=1}^{n}\left(b_{k}-a_{k}\right)=0.

Note 1. We have \overline{x} \in(\overline{a}, \overline{b}) iff the inequalities a_{k}<x_{k}<b_{k} hold simultaneously for all k . This is impossible if a_{k}=b_{k} for some k ; similarly for the inequalities a_{k}<x_{k} \leq b_{k} or a_{k} \leq x_{k}<b_{k}. Thus a degenerate interval is empty, unless it is closed (in which case it contains \overline{a} and \overline{b} at least).

Note 2. In any interval A,

d A=\rho(\overline{a}, \overline{b})=\sqrt{\sum_{k=1}^{n}\left(b_{k}-a_{k}\right)^{2}}=\sqrt{\sum_{k=1}^{n} \ell_{k}^{2}}.

In E^{2}, we can split an interval A into two subintervals P and Q by drawing a line (see Figure 2) . In E^{3}, this is done by a plane orthogonal to one of the axes of the form x_{k}=c\left( see §§4-6, Note 2), with a_{k}<c<b_{k} . In particular, if \right. c=\frac{1}{2}\left(a_{k}+b_{k}\right), the plane bisects the k th edge of A ; and so the k th edge-length of P( and Q) equals \frac{1}{2} \ell_{k}=\frac{1}{2}\left(b_{k}-a_{k}\right) . If A is closed, so is P or Q, depending on our choice. (We may include the "partition" x_{k}=c in P or Q . )^{1}

Now, successively draw n planes x_{k}=c_{k}, \quad c_{k}=\frac{1}{2}\left(a_{k}+b_{k}\right), \quad k=1,2, \ldots, n . The first plane bisects \ell_{j} leaving the other edges of A \mathrm{un}- changed. The resulting two subintervals P and Q then are cut by the plane x_{2}=c_{2}, bisecting the second edge in each of them. Thus we get four subintervals (see Figure 3 for E^{2}. Each successive plane doubles the number of subintervals. After n steps, we thus obtain 2^{n} disjoint intervals, with all edges \ell_{k} bisected. Thus by Note 2, the diagonal of each of them is

\sqrt{\sum_{k=1}^{n}\left(\frac{1}{2} \ell_{k}\right)^{2}}=\frac{1}{2} \sqrt{\sum_{k=1}^{n} \ell_{k}^{2}}=\frac{1}{2} d A.

Screen Shot 2019-05-29 at 11.47.41 PM.png

Note 3. If A is closed then, as noted above, we can make any one (but only one ) of the 2^{n} subintervals closed by properly manipulating each step.

The proof of the following simple corollaries is left to the reader.

Corollary \PageIndex{1}

No distance between two points of an interval A exceeds d A, its diagonal. That is, (\forall \overline{x}, \overline{y} \in A) \rho(\overline{x}, \overline{y}) \leq d A

Corollary \PageIndex{2}

If an interval A contains \overline{p} and \overline{q}, then also L[\overline{p}, \overline{q}] \subseteq A.

corollary \PageIndex{3}

Every nondegenerate interval in E^{n} contains rational points, i.e., points whose coordinates are all rational.

(Hint: Use the density of rationals in E^{1} for each coordinate separately.)


This page titled 3.3: Intervals in Eⁿ is shared under a CC BY 3.0 license and was authored, remixed, and/or curated by Elias Zakon (The Trilla Group (support by Saylor Foundation)) via source content that was edited to the style and standards of the LibreTexts platform.

Support Center

How can we help?