Investigations into the Laplacian and the Laplace
Transform with Respect to the Circle, the Aster,
and Construction of an Analytic Geometric Lens
Jonathan David Evenboer
Philomath, OR
2/14/2016
To fulfill the requirements of an ongoing love affair with mathematicsDedicated to Ms. Mitchell, my 4" & 5" grade math teacher. Thanks for teaching
us to calm our minds before we started doing our arithmetic. It took about 20
years to sink in, but mind calming techniques are now the most important tool in
my mathematical toolbox.February 10, 2016
L Preface/Introduction
In continuation of previous work ((1},[2)), this paper is written with respect to
the properties of the Laplacians and Laplace Transforms of the circle and the
aster and the comparative qualities of these properties. The goal of this paper,
than, is two-fold:
1) To, as stated before, compare qualities and properties of the circle and aster;
2)To further development on the creation of a geometric lens in 2-space, with
an eye on development of a geometric lens in 3-space.
Section 1 will define the Laplacian of a manifold, as well as the Laplacian of
‘a graph. Section 2 will give results of Laplacians with respect to As(x) and
Ci(#)(we note that Laplacians of Sq(*) and Sq/(x) are trivial for our analysis
in 2-space and are otherwise ignored, but for construction of the geometric lens
{in 3space the Euclidean element is absolutely necessary).
Section 3 will be a brief look at the properties of the Laplace Transforms, fol
lowed by analysis specific to Laplace Transforms of the circle and the aster.
Section 4 will conclude the paper, giving final thoughts and setting up fi
‘ture/continued research with respect to the construction of a geometric lens
and other comparative and analytical methods.
Before delving into the more technical aspects of these Laplace-named con-
‘cepts, I'd like to conclude the introduction with a heuristic look at the overar-
‘ching themes this paper (hopefully) constructs and elucidates upon. Thope this
heuristic take sheds more light on why this analysis is interesting from a geo-
‘metric/topological point of view, if not a general mathematical point of view. I,
‘myself, find a good amount of insight into “technical” mathematics when I have
a fairto-good amount of insight into the more intuitive aspects of the particular
mathematics being studied.
Tn regards to Laplacians, we're especially interested in their role with respect to
the heat kernels of a manifold or graph. In this paper we're concerned primarily
‘with the geometric topological representation (manifolds), but wish to empha-
size the equally important graph theoretical representations and applications.
‘We note, via Chung [3], that the definition of a heat kemel on a manifold is,
indeed, dual to the definition of a heat kernel on a graph. In researching this
‘material, I found it beneficial to keep this duality in mind as I moved betweenthe different notations of different disciplines. The power of the heat kernel,
not only in and of itself, but in it’s ability to make interdisciplinary connections
was certainly amazing “discovery” for me. In fact, given the proper amount of
sleepless nights, one can see precisely how these seemingly disparate disciplines
‘can coexist in a fruitful harmony when viewed and studied via the proper tech-
niques.
In regards to Laplace Transforms, the inclusion of this concept is a direct result
of one sleepless night where I mixed up calculations of Laplacians and Laplace
‘Transforms. Despite the initial error, a few fruitful facts and ideas resulted
(which just goes to show: never ignore accidents, mistakes, errors, or things
otherwise known as “blunders”. The main result, which we explore in this
paper, is the usefulness of the aster/hyperbolic octahedron as a complex-valued
analogue for the circle/sphere. The usefulness comes from the fact that while
‘the Laplace Transform of circle/sphere is asymptotic as it approaches the geo-
metric poles, the aster /hyperbolic octahedron has no asymptotes at any poles
(im fact, all poles have rational number valves).
Annote on references: other works not direetly referenced in this work, but which
T consider necessary (if not enjoyable) supplemental material, will be listed in
the bibliography after the directly referenced resources. Also given is a list of
basic/foundational notations, which will hopefully provide the reader with an
increased ease of reference with respect to my sometimes tragically non-standard
notations. Some figures (eg; images via Wolfram or Mathlets) will be referenced
Jn the work, but will be located in an appendix of figures. This is & direct result
of my ignorance of public vs. private domain in regards to mathematical images
‘on the internet.
2 Section 1: the Laplacian
We begin by giving two dual definitions for the Laplacian: the differential and
sraph theoretical definitions, respectively.
Def.l (differential)[4] : Any = — Zr, where z is some arbitrary variable and the
second derivative is ofthe standard definition [eale,iff geo texts
Def.2 (graph theoretical) (3]AG=(1 if u = and dy #0; —1//dydy if wis adja-
cent to ¥; 0 in all other instances }
Other than differences in notation, both definitions otherwise obey the stan-
dard, clementary properties of the Laplacian. We note that Def. 1 is simply the
2nd (partial) derivative, which gives a metric on a Reimannian manifold (ref:
iff geo calc texts). It’s of note that Def. 2 has a dual expression
Ac =1-(1/k)A
ifG is k-regular [3]. Since the 1 and 2 manifolds being studied ae all kregular
in their graph theoretical representations, we will adopt the k-regular notation
for the remainder of the paper (and forthcoming work dealing with graph theo-
retical aspects) noting that this implies Ac is the eigenvalue decomposition of‘the matrix representation of some k-regular graph G (3)
3 Section 2: Results with respect to As, Ci
Some specific results with respect to Def. 1 are given in this section. Specifically
we're interested in comparative results of Ci(s) and As(«).
AnCi(s)
Ay Aa(«) = (—8{sin(t) — dsin(Bt) + cos(t) + 8cos(3t)))/(4r?)
Note that Ass Ci(#) is everywhere trivial, while Ay As(s) allows for non-
zero results. Recall from [2] that for As(*)
1 = |sec6|(1 + tan?/*6)?/?
With respect to it’s role in heat kernel ealculations, the Laplacian of As(+) gives
the following results
RAs) = Ayu(As)(cos*(@) + sin¥(0))?
noting that A(Ci) = 0 everywhere. In this respect, though the two mani-
folds are topologically equivalent, they give much different results in a differen-
tial/geometric sense. When considering methods from [2], the natural question
arises: Can we associate any asymptotic points of Aye As(«) with points on Ay
Ci(x) = 0 such that Ay Ci(s)XAacAs(e) is a smooth, compact, continous
fanction. This may (or may not) give interesting results for a heat kernel influ-
enced by a Laplacian Ane Ci(s) XA As)
4 Section
Ci and As
‘Applying the usual, standard definition of the Laplace Transform [elementary
texts], we have the basic formula
L108) = Fs)
where s € C and (t, f(t) €R For Ci(s) and As(+)we retrieve the following:
£,(Ci) /s
Li(As) = (3° + 78 + 6)/(s4 + 108? + 9)
‘At first glance, the reader will be compelled to calculate the cardinality of the
trivialty of these results. The power of these results lies inthe topological equiv-
lence of the two given values. Indeed, both converge to zero as s approaches
infinity so that
Timy 500 Le(Ci) = lim,o0£1(As) = 0 Where £,(Ci) and £,(As) differ is in
ther exiatence at 5=0:
lim 10 £:(Ci
lim,-+0 £2(As)
No methods of approximation are needed in terms of L4(As), as it exists at
5 ~ 0, noting s —0 is analogous to polar values
8= (0.n2,m,3n2,27}
Laplace Transforms with respect toabe
As(s
‘The critical Laplace Transform points are highlighted in the above figure as
‘marked points on their given polar representations. The key here les in noting
that all highlighted points are equivalent, which is to say
Cilr,8) = As(r,8) |r = {1,0 = {0,/2,n, (8x) /2,2n}
‘The true value of the aster hyperbolic octahedron with respect to the Laplace
‘Transform les in the fact that itis everywhere smooth, compact, complete, and
continuous under these manifolds. We recall from [1] and [2] thet there exists
linear, invertible bijective mappings
(Ore
FS arbitracy)
Ge Ae)
Os [1s . f Mae fil boas
ey ancl
se fae tS
and so any and all information taken from data sets on As(s) has analogous
analytical meaning in Ci(s), Sq(x), Sq(s), as well as their respective analogues
in 3-space ({1}, [2}) when we consider the topological equivalence of these man-
4
wlan (Ash fs teifolds (both geometric and Reimannian) in variables (¢),(t,) (r,8),7(0, 9):
5 Section 4:Conclusion
‘As with much of this paper, our conclusions are fairly concise and seemingly
trivial on the surface of things.
Given topologically equivalent I-manifolds (in our case a “spherical circle” and
4 “hyperbolic circle”), we can retrieve results with more meaning (in terms of
interesting, non-trivial results) when we choose the hyperbolic circle (ie; the
aster, and in $-space its analogue the hyperbolic octahedron),
‘With respect to the Laplacian, and by extension the heat kernel,
AnCi(s)
Aw As(+) = (Afsin(t) — Bsin(3t) + cos(t) + 3cos(3t)))/(4r2)
showing that As(s) gives far more interesting results in both a geometric sense
and a graph theoretical sense (again, invoking duality of the Laplacians of
Reimannian manifolds M and their graph theoretical representation G.
With respect to the Laplace transforms, we note simply that using the hyper
bolic octahedron and aster in lien of the circle or square (and their 3-space
analogues) allows us to use the Laplace transform to retrieve calculations of
‘8 smooth, continuous type on manifolds that would otherwise retrieve results
that are asymptotic in nature (see Appendix of Images for comparative visual-
izations of these Laplace Transforms). ‘The Laplace Transform is easily invert-
‘ble in all cases covered in this paper, and thus all information retrieved from
the Laplace Transforms of the aster/hyperbolic octahedron is easily applied to
Ci(x), Sa(x), Sq'(x) and their analogues in 3-space when the methodology cov-
cred in [2] is applied
References
[1] The Gauss-Bonet Formula: Topological Invariance”, Jonathan D. Evenboer
Fall’12, Oregon State University
2Geometric Analytical Methods In Regards to the Topology of 1-Manifolds of
Constant Curvature, Jonathan D. Evenboer, Spring 2015, Oregon State Univer-
sity (this paper is a direct continuation of the previous work referenced here)
3 Spectral Graph Theory, Fan Chung, AMS (ch.’s 1 and 10 are the chapters
relevant to this paper and future work)
4 The Laplacian on a Riemannian Manifold, Steven Rosenberg, Cambridge Uni-
versity Press (ch. 1 is relevant to this work, ch. 3 is indizeetly relevant to this
‘work but directly relevant to future worl)
5 Differential Geometry Elementary Caleulus/ODE/Linear Algebra Textbooks
(s0¢ references at the end of (2] for relevant texts)