Sie sind auf Seite 1von 14

On An Application of the Geometric Lens gLn(?

):
A Conditional Solution to the Sharp Corner
Problem

Jonathan David Evenboer

Oregon State University

19 July 2020

1 Introduction

In this paper we will explore a practical problem-solving application of the con-


struction gLn (?) = Asn (?) ∪ Cin (?) ∪ Sq 0n (?) developed in [1,2,3,4,5,6,7,8,9]. Via the
construction gLn (?), we will present a conditional solution to the problem of defining
a derivative of the absolute value function at any point 0 in the domain (henceforth
referred to as “the sharp corner problem”).

In Section 1, we will review the construction gLn (?). If the reader is at all fa-
miliar with the previous work in developing gLn (?) in the sources cited above, they
will notice some adjustments in the definition of the construction itself. Namely,

1
we now define Sq 0n (?) by the linear equation where each term in the summation is
under the absolute value function. For example, in Cartesian co-ordinates in Rn , we
define Sq 0n (?) = |x1 | + |x2 | + . . . . + |xn+1 |, as opposed to using the too-handwavy
description “the square/cube/n-cube rotated”. In general, attempts were made to
refine previous development of gLn (?) and to correct mistakes that were addressed
in errata of [1,2,3,4,5,6,7,8,9]. The same reader will notice that, here, we use position
vector notation while still referring to the components of gLn (?) as n-manifolds. For
this particular problem, the position vector notation is wieldier and more appropriate
to the problem. The authour also admits that the use of the phrase ’n-manifold’ is
probably problematic here. The authour was striving for continuity in terminology
in previous papers, but acknowledges that in the context of this paper, it is prob-
ably better to refer to gln (?) and its component curves and surfaces as ’geometric
objects’ as opposed to ’n-manifolds’. Also, the term ’construction’ has been adopted
for gLn (?) in this particular paper.

Section 2 builds upon Section 1, culminating in a theorem offering a conditional


solution to the sharp corner problem. After presentation of the theorem, we will
discuss some of the short-comings of this conditional solution. Specifically, we will
address the fact that, while the conditional solution does indeed assign a defined
derivative at every element of the domain, we still end up with a jump discontinuity
when taking the derivative in the negative direction.

2
2 Section 1: The Geometric Lens gLn(?)

Before we can present the theorem central to this paper, we must first set up the
necessary construction.
The construction of the Geometric Lens gLn (?) = Asn (?) ∪ Cin (?) ∪ Sq 0n (?) is
absolutely necessary for the solution we present (the term “conditional solution” will
be frequently used to emphasize this dependence). We begin, then, by defining the
components comprising gLn (?), starting with ?.

Def. 1: ?
? = (?1 , ?2 , . . . , ?n+1 ) is a variable set acting as the domain for an implicit expression
of an n-manifold X n (?).

From here, we will define the specific n-manifolds X n (?) we will be using in our
construction of gLn (?).

Def. 2: Sq 0n (?)
Sq 0n (?) = Sq 0n (?1 ) + Sq 0n (?2 ) + . . . + Sq 0n (?n+1 )is the implicit expression for the unit
n-octahedron.

For example, if ? = (x, y) and Sq 01 (x, y) ⊆ R2 , then Sq 0n (x, y) = |x| + |y|. If


? = (1, θ) and Sq 01 (1, θ) ⊆ C2 , then Sq 01 (1, θ) = | cos θ| + |i sin θ|. So on and so forth.

3
Def. 3: Cin (?)
Cin (?) = Cin (?1 ) + Cin (?2 ) + . . . + Cin (?n+1 ) is the implicit expression for the unit
n-sphere.

For example, if ? = (x, y) and Ci1 (x, y) ⊆ R2 , then Ci1 (x, y) = x + y. If ? = (1, θ)
and Ci1 (1, θ) ⊆ C2 , then Ci1 (1, θ) = cos θ + i sin θ. So on and so forth.

Def. 4: Asn (?)


Asn (?) = Asn (?1 ) + Asn (?2 ) + . . . . + Asn (?n+1 ) is the unit n-hyperbolic octahedron.

2 2
For example, if ? = (x, y) and As1 (?) ⊆ R2 , then As1 (x, y) = x 3 + y 3 . If ? = (1, θ)
and As1 (1, θ) ⊆ C2 , then As1 (1, θ) = cos3 θ + i sin3 θ. So on and so forth.
With the n-manifold components of gLn (?) defined, we turn to the actual construc-
tion of gLn (?) itself. We begin by defining the set of the corner points of X n (?).

Def. 5: {CP }X n (?)


For an arbitrary X n (?) ∈ {Asn (?), Cin (?), Sq 0n (?)}, we define
{CP }X n (?) = {Asn (?) ∩ Cin (?) ∩ Sq 0n (?)}.

For example, where ? = (x, y) with X n ⊆ R2 ,


{CP }Asn (?) = {CP }Cin (?) = {CP }Sq0n (?) = {(1, 0), (0, 1), (0, −1), (−1, 0)}.
The preceding definitions give us the necessary machinery for construction of
gLn (?) = Asn (?) ∪ Cin (?) ∪ Sq 0n (?).

4
Proposition 1
If the union gLn (?) = Asn (?) ∪ Cin (?) ∪ Sq 0n (?) exists such that
{Asn (?) ∩ Cin (?) ∩ Sq 0n (?)} = {CP }gLn (?) , then gLn (?) = Asn (?) ∪ Cin (?) ∪ Sq 0n (?)
(such that {CP }gLn (?) = {CP }Cin (?) ) is a constructible structure.

Proof
Centre Asn (?) and Sq 0n (?) in Cin (?) with respect to the centre of Cin (?) so that
{Asn (?)∩Cin (?)∩Sq 0n (?)} = {CP }Cin (?) = {CP }Asn (?) = {CP }Sq0n (?) . Then, taking
the union gLn (?) = Asn (?) ∪ Cin (?) ∪ Sq 0n (?), we have
{CP }gLn (?) = {CP }Cin (?) = {CP }Asn (?) = {CP }Sq0n (?) .
Under the equality, we can choose {CP }gLn (?) = {CP }Cin (?) . Thus we have that, if
the union gLn (?) = Asn (?) ∪ Cin (?) ∪ Sq 0n (?) exists such that
{Asn (?) ∩ Cin (?) ∩ Sq 0n (?)} = {CP }gLn (?) , then
gLn (?) = Asn (?) ∪ Cin (?) ∪ Sq 0n (?) such that {CP }gLn (?) = {CP }Cin (?) is a con-
structible structure. 

Further subtleties resulting from this construction are necessary to prove the theo-
rem central to this paper, but the above gives a sufficient description of the actual
structure gLn (?) and our ability to choose {CP }gLn (?) = {CP }Cin (?) . The ability to
choose {CP }gLn (?) = {CP }Cin (?) is a crucial step in reaching our conditional solution
to the sharp corner problem.

5
3 Section 2: A Conditional Solution to the Sharp

Corner Problem

It is a well-known fact that the derivative of the absolute value function is undefined
at zero. As mentioned earlier, this is commonly referred to as the “sharp corner
problem”, a term we will use for the remainder of this paper.
In this section, we will use the construction gLn (?) developed in the previous section
(as well as [1,2,3,4,5,6,7,8,9]) to give a possible solution to this problem. As the title
of this section (and of this paper) suggest, this potential solution is dependent on
the utilization of properties of gLn (?) = Asn (?) ∪ Cin (?) ∪ Sq 0n (?) where
{CP }gLn (?) = {CP }Cin (?) = {CP }Asn (?) = {CP }Sq0n (?) .

Recall from Def. 2 that Sq 0n (?) is expressed as the sum of n + 1 variables, with
each term under the absolute value function. For example, in R2 , Sq 01 (?) = |x| + |y|
for ? = (x, y), where ∂
∂x
Sq 01 (x, y) = ∂
∂x
|x| +0 = x
|x|
such that x 6= 0. Likewise,
y

∂y
Sq 01 (x, y) = ∂
∂y
|y| +0 = |y|
such that y 6= 0. In C2 the partial derivative with
y
respect to x is the same as the real case. In fact, ∂
∂y
Sq 01 (x, iy) = ∂
∂y
|iy| + 0 = |y|
such
that y 6= 0 exactly as in the real case.

Since the partial derivatives are the same for Sq 0n (?) ⊆ Rn+1 and Sq 0n (?) ⊆ Cn+1 ,
the complex case will often be referred to in passing after treatment of the real case.

6
Observe that in both Rn+1 and Cn+1 , the partial derivatives of Sq 0n (?) are undefined
exactly at {CP }Sq0n (?) , a finite set.

The goal of this section, then, is to find a way to induce a defined derivative at
the points in {CP }Sq0n (?) by exploiting the fact that
{CP }Sq0n (?) = {CP }Cin (?) ⊆ gLn (?).
Before we can show the above, we must introduce a few more substructures and
properties of gLn (?).

Def. 6: Op(X n (?))


For any n-manifold X n (?) ⊆ gLn (?), Op(X n (?)) = X n (?) \ {CP }gLn (?) where
{CP }gLn (?) = {CP }X n (?) by definition.

For example, in R2 we have Op(As1 (?)) = Asn (?) \ {CP }gL1 (?) .
Since {CP }gLn (?) is a set of singletons such that CPi ∩CPj = ∅, all CPi ∈ {CP }gLn (?)
are closed. Then, for any arbitrary X n (?) ⊆ gLn (?), Op(X n (?)) is an open set for
all n ∈ N.
In fact, def. 6 implies that, for X n (?), Y n (?), Z n (?) ⊆ gLn (?), we can preform the
following operation:

gLn (?) \ Op(X n (?)) = (X n (?) ∪ Y n (?) ∪ Z n (?)) \ Op(X n (?))


(1)
n n
= Y (?) ∪ Z (?)

7
where {CP }gLn (?) ⊆ gLn (?) \ Op(X n (?)). This leads to the following proposition:

Proposition 2
If gLn (?) = X n (?) ∪ Y n (?) ∪ Z n (?), and
X n (?) ∪ Y n (?) ∪ Z n (?) = Op(X n (?)) ∪ Op(Y n (?)) ∪ Op(Z n (?)) ∪ {CP }gLn (?) ,
then gLn (?) \ {Op(X n (?)), Op(Y n (?))} = Op(Z n (?) ∪ {CP }gLn (?) = Z n (?), where
{CP }gLn (?) = {CP }Cin (?) is a connected, continuous construction.

Proof
For X n (?), Y n (?), Z n (?) arbitrary elements of {Cin (?), Asn (?), Sq 0n (?)}, assume
gLn (?) = X n (?) ∪ Y n (?) ∪ Z n (?) where
{CP }gLn (?) = {CP }X n (?) = {CP }Y n (?) = {CP }Z n (?) . Since Cin ∈ {X n (?), Y n (?), Z n (?)},
we can choose {CP }gLn (?) = {CP }Cin (?) , by the proof of Proposition 1.
By def. 6, we have

gLn (?) = X n (?) ∪ Y n (?) ∪ Z n (?)


(2)
n n n
= Op(X (?)) ∪ Op(Y (?)) ∪ Op(Z (?)) ∪ {CP }Cin (?)

where gLn (?) is closed, connected, and continuous.


Then, taking gLn (?) \ Op(X n (?)), we can see that

gLn (?) \ Op(X n (?)) = Op(Y n (?)) ∪ Op(Z n (?)) ∪ {CP }Cin (?)
(3)
= Y n (?) ∪ Z n (?)

8
is also a closed, connected, and continuous construction.
A second iteration of the above process gives

gLn (?) \ {Op(X n (?)), Op(Y n (?))} = Op(Z n (?)) ∪ {CP }Cin (?)
(4)
= Z n (?)

where we can choose {CP }Cin (?) = {CP }Z n (?) regardless of our choice of Z n (?).
Since Z n (?) is in {Cin (?), Asn (?), Sq 0n (?)} = A, and every element of A is closed,
connected, and continuous, we must have that the construction

gLn (?) \ {Op(X n (?)), Op(Y n (?))} = Op(Z n (?)) ∪ {CP }Cin (?)
(5)
n
= Z (?)

is also closed, connected, and continuous. 

With the proof of Proposition 2, we finally have the last piece in place to give
(and prove) the central theorem of this paper.

Theorem
If gLn = Asn (?) ∪ Cin (?) ∪ Sq 0n (?) with
{CP }gLn (?) = {Asn (?) ∩ Cin (?) ∩ Sq 0n (?)} = {CP }Cin (?) ,
then gLn (?) \ {Op(Cin (?)), Op(Asn (?))} = Op(Sq 0n (?)) ∪ {CP }Cin (?) = Sq 0n (?), and
Sq 0n (?) is differentiable at all points on domain ?.

9
Proof
Let {X n (?), Y n (?), Z n (?)} = {Cin (?), Asn (?), Sq 0n (?)}.
By Proposition 2,

gLn (?) = X n (?) ∪ Y n (?) ∪ Z n (?)

= Cin (?) ∪ Asn (?) ∪ Sq 0n (?)


(6)
n n 0n
= Op(Ci (?)) ∪ Op(As (?)) ∪ Op(Sq (?)) ∪ {CP }gLn (?)

= Op(Cin (?)) ∪ Op(Asn (?)) ∪ Op(Sq 0n (?)) ∪ {CP }Cin (?)

where

gLn (?) \ {Op(Cin (?)), Op(Asn (?))} = Op(Sq 0n (?)) ∪ {CP }Cin (?)
(7)
0n
= Sq (?)
.
Then, for all ?i ∈ (?1 , ?2 , . . . , ?n+1 ) = ? ⊆ Rn+1 (or Cn+1 ), we have
lim Sq 0n (?) = lim Op(Sq 0n (?)) ∪ lim {CP }Cin (?) .
?i →∞ ?i →∞ ?i →∞

Taking partial derivatives, we have

10
∂ ∂ ∂
lim Sq 0n (?) = lim Op(Sq 0n (?)) ∪ lim {CP }Cin (?)
?i →∞ ∂?i ?i →∞ ∂?i ?i →∞ ∂?i
∂ ∂
= lim Op( Sq 0n (?)) ∪ lim {CP }Cin (?)
?i →∞ ∂?i ?i →∞ ∂?i

= {−1, 1} ∪ {1}

= {−1, 1}
∂ ∂ ∂
= { lim − Op( Sq 0n (?)), lim + Op( Sq 0n (?)), lim {CP }Cin (?) }
?i →∞ ∂?i ?i →∞ ∂?i ?i →∞ ∂?i

(8)

∂ ∂
where lim + ∂?i
{CP }Cin (?) = lim − ∂?i
{CP }Cin (?) = 1, and where
?i →∞ ?i →∞
∂ 0n
∂?i
Op(Sq (?j )) = 0 when i 6= j.
Thus, under the geometric lens gLn (?), if we take

gLn (?) \ {Op(Asn (?)), Op(Cin (?))} = Op(Sq 0n (?)) ∪ {CP }Cin (?)
(9)
= Sq 0n (?)

then we have that lim + ∂


∂?i
Sq 0n (?) = 1, including at ?i = 0, and
?i →∞
lim − ∂
∂?i
Sq 0n (?) = −1, excluding the point associated with ?i = 0, where the limit
?i →∞
of the partial derivative is still one even in the negative direction, and so the deriva-
tive of the absolute value function can be defined at every point in the domain ?,
including ?i = 0 ∈ ? in both Rn+1 and Cn+1 . 

The following example illustrates the result of the preceding theorem.


Let gL1 (?) \ {Op(Ci1 (?)), Op(As1 (?))} = Op(Sq 01 (?)) ∪ {CP }Ci1 (?) where

11
? = (x, y) ⊆ R2 .
Then ∂
∂x
Op(Sq 01 (x, y)) = x
|x|
for all x 6= 0 and ∂
∂x
{CP }Ci1 (x,y) = 1 for all
x ∈ {CP }Ci1 (?) .
From here, we can see that
lim x = −1, and lim+ x
= 1 for all x ∈ Op(Sq 0n (?)),
x→∞− |x| x→∞ |x|

and that lim− 1 = lim+ 1 = 1 when x ∈ {CP }Cin (?) . The same can be shown for
x→∞ x→∞
01
all y ∈ Op(Sq (x, y)) ∪ {CP }Cin (?) We can easily extend the above to all ? ⊆ Rn+1 .
In Cn+1 , the process is the same, with the result differing only in the complex vari-
able.
y
For example, in C2 , we still have that ∂
∂y
Op(Sq 01 (x, iy)) = |y|
, but we have that

∂y
{CP }Ci1 (x,iy) = i, with the limit obviously going to i in both the positive and neg-
ative directions. Partial differentiation with respect to the x variable in {CP }Ci1 (x,iy)
is the same as in the R2 case.

The above examples, as well as the proof of the theorem, highlight the conditional
nature of the result, which is completely dependent on the construction gLn (?).
While the above theorem does allow for the existence of the derivative at zero in the
domain of the absolute value function (in real or complex variables), we have also
seen this derivative is not continuous in the negative direction, since

∂ ∂ ∂
lim− Sq 0n (?) = Op( lim− Sq 0n (?)) ∪ lim− {CP }Cin (?)
?→∞ ∂?i ?→∞ ∂?i ?→∞ ∂?i

= {−1} ∪ {1} (10)

= {−1, 1}

12
implies there exists a jump discontinuity at all CPi ∈ {CP }Cin (?) .
While the theorem is specifically written for the Sq 0n (?) case, it’s also true that the
same logic can be applied to the Asn (?), which is representative of the family of
manifolds/geometric objects with cuspoid points. The process is exactly the same,
only swapping out Sq 0n (?) with Asn (?). Unfortunately, the jump discontinuity in
the negative direction also remains.
In both the absolute value and cuspoid cases, then, the derivative may be defined at
each point, but it is necessarily pointwise.
Still, it is interesting to observe that the theorem applied to Sq 0n (?) gives a second
partial derivative that is zero everywhere on Sq 0n (?) , as well as at all
CPi ∈ {CP }Asn (?) ⊆ Asn (?), as claimed in [1,2,3,4,5,6,7,8,9].

References
1. The Gauss-Bonnet Formula: Topological Invariance
(https://jonathandavidevenboer.weebly.com/blog/1st-paper)
2. Geometric Analytical Methods In Regards to the Topology of 1-Manifolds of Con-
stant Curvature
(https://jonathandavidevenboer.weebly.com/blog/new-paper)
3. Investigations Into Laplacians and Laplace Transforms With Respect to the Cir-
cle, the Aster, and the Construction of an Analytic Geometric Lens
(https://jonathandavidevenboer.weebly.com/blog/3rd-papaer)

13
4. Determinants and Applicable Methodology of Determinants with Respect to the
Building and Understanding of Geometric Lens Theory
(https://jonathandavidevenboer.weebly.com/blog/fourth-paper)
5. Concentrations of Positive Curvature on Manifolds with Euler Characteristic 2
for Real Dimensions 3 and Higher
(https://jonathandavidevenboer.weebly.com/blog/new-math-paper-5)
6. Aligning Normals In a Low Dimensional Geometric Lens Space
(https://jonathandavidevenboer.weebly.com/blog/new-geometric-lens-related-paper)
7. A Composition of Functions With Respect to the Justification of the Claim that
As1 (?) is the Hyperbolic Element of gL1 (?)
(https://jonathandavidevenboer.weebly.com/blog/this-years-math-paper-regarding-geometric-
lens-stuff)
8. Errata for Multiple Papers, Part 1
(https://jonathandavidevenboer.weebly.com/blog/major-errata-alert)
9. Errata for Multiple Papers, Part 2
(https://jonathandavidevenboer.weebly.com/blog/ultra-major-errata-alert-times-20-
to-the-8th-power-or-something)

14

Das könnte Ihnen auch gefallen