Two-Dimensional Geometries and Topologies, generated by a ...
Transcript of Two-Dimensional Geometries and Topologies, generated by a ...
Two-Dimensional Geometries and Topologies, generated by a special
inner product for Complex-Type Numbers
A brief report on work performed in the topic for evaluation process
by
Yellanki Abhinav
(08MA2013)
Under the guidance of
Prof. Debapriya Biswas
Department of Mathematics
Indian Institute of Technology, Kharagpur
2
CERTIFICATE
This is to certify that the M.Sc. project entitled Two-Dimensional and higher
Dimensional Geometries, generated by a special inner product for Complex-Type
Numbers submitted by Yellanki Abhinav (08MA2013) to the Department of
Mathematics in partial fulfillment of the requirements for the degree of Master of
Science is a bonafide record of the work carried out by him under my supervision
and guidance during the academic year 2012-2013.
Professor Debapriya Biswas Department of Mathematics
Indian Institute of Technology, Kharagpur
Kharagpur 721302
India
Date:
Signature:
3
Acknowledgement
I declare all that this report is my own work. And I would like to express my
sincere gratitude to Professor Debapriya Biswas for providing me with the
opportunity to work in this project under her supervision. Her able guidance and
constant inspiration was a major source of motivation to help me carry out this
semester project.
Abhinav Yellanki
(08MA2013)
4
TABLE OF CONTENTS:
Sl.no Topic Page No
1. Abstractβ¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦. 5
2. Introductionβ¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦ 6
3. Preliminariesβ¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦.. 6
4. General complex type number systems β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦. 7
5. Geometries induced on πΉπby π β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦. 8
6. Topologies induced on πΉπby π β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦.. 16
7. Futurescope β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦. 18
Referencesβ¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦β¦........... 18
5
1. Abstract
In non relativistic classical mechanics we use Euclidean space[1][2], as time is treated as
independent of state of motion of the observer and the object. which wonβt work in relativistic
mechanics since they usually have a time dimension. Theory of Relativity is a theory of Spacetime.
Spacetime is Mathematical model that combines space and time into a single continuum. It is usually
interpreted as 1or 2or 3 spatial dimensions along with a time dimension which isnβt spatial. Spacetime
can be viewed as the consequence of Albert Einsteinβs βtheory of special relativityβ[1] It is explicitly
proposed mathematically by the mathematician Hermann Minkowski in a essay, building on and
extending Einsteinsβs works.
In his work on special theory of relativity [3], G.L Naber provides a mathematically rigorous way
to understand the special theory of relativity showing the physical significance of mathematics. His model
is based on a special indefinite inner product.
Paul Fjelstad has worked on this field extending Naberβs work by showing that inner product and
topologies from Naberβs work can be generated by the hyperbolic complex numbers[4]. He studied two
dimensional geometries and physics generated in a similar manner to Naber for more general Complex
Type Numbers πΆπ = { π§ = π₯ + ππ¦; π₯,π¦ β πΉ, π Ι πΉ, π2 = π΄ + π 2π΅ π€ππππ π΄,π΅ β πΉ } by defining a
more general inner product. His work starts with remark that all the general Complex type numbers πΆπ are
ring isomorphic to one of the three types of familiar complex type numbers(elliptic, hyperbolic and dual).
Later he proves these complex type number systems are as a group of homeomorphisms to each other on
the topologies induced by the more general inner product he constructed.
Paul Fjelstadβs paper contain a number of propositions which are drawn from the similarities with
the Naberβs work on Minkowski space (π΄ with three spacial dimensions and a time dimension). In this
paper the main focus is on understanding, verifying and proving the various propositions, lemmas and
theorems in the works mentioned above.
6
2. Introduction
G.L Naber model is based on a special indefinite inner product π1:π΄ x π΄ β πΉ [3] and defined as
π1 π£,π€ = π£1π€1 + π£2π€2 + π£3π€3 β π£4π€4
where π£,π€ β π 4 and π£1 ,π£2 ,π£3 , π£4 ,π€1 ,π€2 ,π€3 ,π€4 β π
π1πs of index β1β and the associated Lorentz group of transformations in π 4. Index indicate
negative definiteness of the inner product it is otherwise the dimension of subspace of a given vector
space for which π(π£, π£) = 0 where π£ β ππππ‘ππ πππππ on which π is defined. His book also presents
some collection of unique work in topologies of the 4-D geometry he worked upon.
Paul Fjelstad defined inner product for more general Complex Type Numbers[4] πΆπ = { π§ = π₯ +
ππ¦; π₯, π¦ β πΉ, π Ι πΉ, π2 = π΄ + π 2π΅ π€ππππ π΄,π΅ β πΉ } as π: πΆππ΄,π΅ π₯ πΆπ
π΄,π΅ β πΉ and π π£,π€ =
π£π€ β² + π€π£ β² /2 Where π§ β πΆπ , π§ = π₯ + ππ¦ and the conjugate is defined as π§β² = π₯ + 2π΅π¦ β ππ¦. Which can be
explicitly written as
π π£,π€ = π£1π€1 + π΅ π£1π€2 + π£2π€1 β π΄π£2π€2 (1)
Where π£,π€ β πΆππ΄,π΅
and π£1 ,π£2 ,π€1 ,π€2 β πΉ
Which is also of Index β1β and gives Naber inner product when B=0 and A=1. Likewise most of
the geometrical properties induced by inner product are also similar to that of the properties in Naberβs
works.
3. Preliminaries Indefinite Inner product:
This type of inner product satisfies all properties of standard inner product except the positive
definiteness property.
i.e., g(x,x) β₯ 0 is not always true.
Index of Inner product:
This is defined as the number of basis vectors of a finite dimensional vector space V with orthonormal
basis is of the form {e1, e2, β¦ en}, for which g(ei, ei) β€ 0 where i Ο΅ {1,2,..n}.
Null vector, Null Cone, Null wordline:
7
π₯ β π 2 will be called null (or light like) if πβ²(π₯, π₯) = π. The null (light) cone πΆπ(π₯0) at π₯0 β π 2 is
defined by πΆπ(π₯0) = {π₯ β π 2; πβ²(π₯ β π₯0) = 0}, where πβ(π₯) = πβ(π₯, π₯). For π₯ β πΆπ π₯0 ,π₯ β π₯0
the null wordline (of light ray) containing π₯0 and π₯ is defined by
πΉππ,π = {π = ππ + π(π β ππ); π β πΉ}
Spacelike, Timelike:
If πβ(π₯) > 0, π₯ will be called spacelike and if πβ²(π₯) < π, x will be called timelike.
Collection of all timelike vectors be π.
Time cone:
For each timelike π₯0 we define the time cone πΆπ(π₯0), future time cone πΆπ+(π₯0) and past time cone πΆπ
β π₯0
at π₯0 by
πΆπ(π₯0) = {π₯ β π 2; πβ² π₯ β π₯0 < 0}
πΆπ+ π₯0 = πΆπ π₯0 β© π+, πΆπ
β π₯0 = πΆπ π₯0 β© πβ
Duration of π₯:
For any π₯ β π one defines the duration of π₯ by πβ (x) = βπΈβ²(π) .
Timelike straight line:
A subset of π 2of the form π₯0 + π‘ π₯ β π₯0 ; π‘ β π where πβ²(π₯ β π₯0) < 0, is called timelike straight
line in π 2.
4. General complex type number systems
A general complex type number system forms an Algebraic ring and can be written in the form
π§ = π₯ + ππ¦ where π₯,π¦ π π and π Ι π Satisfies the equation π2 = π΄ + π 2π΅ with fixedπ΄,π΅ π πΉ. In
general all the complex type number systems can be represented by Cq = {z = x + qy; x, y Ο΅ R }. All the
number systems of Cq are ring isomorphic with one of the three types of familiar complex type numbers
1. System of Complex numbers, if π΅2 + π΄ < 0
2. System of Dual numbers, if π΅2 + π΄ = 0
3. System of hyperbolic Complex numbers, if π΅2 + π΄ > 0
To prove the ring isomorphism, we can construct a function which establishes the ring isomorphism
present. In the case of B2 + A < 0 we can use the function β : πΆππ΄,π΅ β πΆπ
1,0 defined as
β π₯ + ππ¦ = π₯ + π΅π¦ + π π΄ + π΅2
Where π2 = 1, π2 = π΄ + π(2π΅) and π΅2 + π΄ > 0
8
So inverse function will be
β β1 π₯ + ππ¦ = π₯ βπ΅π¦
π΄ + π΅2 + π π΄ + π΅2 π¦
As the inverse exists β is ontoβ¦..(1)
Consider π§1 , π§2 β πΆππ΄,π΅
where π§1 = π₯1 + π¦1 πππ π§2 = π₯2 + π¦2
Now let us check for one-one
Let β π§1 = β π§2 β π₯1 + ππ¦1 = β π₯2 + ππ¦2
π₯1 + π΅π¦1 + π π΄ + π΅2 π¦1 = π₯2 + π΅π¦2 + π π΄ + π΅2 π¦2
As real and imaginary parts have to be equal
π΄ + π΅2 π¦1 = π΄ + π΅2 π¦2 & π₯1 + π΅π¦1 = π₯2 + π΅π¦2
π¦1 = π¦2 & π₯1 = π₯2 Which implies π₯1 + ππ¦1 = π₯2 + ππ¦2 or π§1 = π§2 and β is one-one β¦.(2)
From hence (1) & (2) β is bijective.
Now β (π§1 +π§2) = β ( (π₯1 + π₯2) + π(π¦1 + π¦2))
= (π₯1 + π₯2) + π΅ π¦1 + π¦2 + π π΄ + π΅2 π¦1 + π¦2
= π₯1 + π΅π¦1 + π π΄ + π΅2 π¦1 + π₯2 + π΅π¦2 + π π΄ + π΅2 π¦2 β (π§1 +π§2) = β (π§1) + β (π§2) β (π§1 π§2) = β ((π₯1 + ππ¦1) π₯2 + ππ¦2 )
= β ((π₯1π₯2 + π΄π¦1π¦2 ) + π π₯1π¦2 + π¦1π₯2 + 2π΅π¦1π¦2 ) = π₯1π₯2 + π΄π¦1π¦2 + π₯1π¦2 + π¦1π₯2 + 2π΅π¦1π¦2
+ π π΄ + π΅2 π₯1π¦2 + π¦1π₯2 + 2π΅π¦1π¦2
= π₯1 + π΅π¦1 + π π΄ + π΅2 π¦1 . π₯2 + π΅π¦2 + π π΄ + π΅2 π¦2
= β π₯1 + ππ¦1 .β π₯2 + ππ¦2 = β (π§1) . β (π§2) hence β is ring isomorphism Likewise we can use the same function for establishing ring isomorphism in the remaining two cases too.
5. Geometry Induced on πΉπ by π
Few important properties of π given by (1)
Lemma:
(i) For all π΄,π΅ β π the function π in (i) is bilinear and symmetric.
(ii) π is nondegenerate (i.e. π(π£,π€) = π, β π€ β π 2 implies v = (0, 0)) if and only if π΅2 + π΄ β π
(iii) If π΅2 + π΄ > 0 then π is of index 1, if π΅2 + π΄ < 0 then π is of index 0.
9
Lemma (iii) implies that indefiniteness of inner product in the systems of complex type numbers
is exclusively due to number systems which are ring isomorphic with hyperbolic complex numbers. And
other two types of number systems are usual positive definite and follow all geometrical properties of
definite inner product spaces.
So, we study only systems of complex type numbers satisfying π΅2 + π΄ > 0. For clarity let the
inner product π in (1) be denoted by πβ². The above lemma permits us to extend all the results of Naberβs
book to number systems for which π΅2 + π΄ > 0. Few of the important results are mentioned below
Theorem (1): Two nonzero null vectors π₯,π¦ β π 2 are πβ-orthogonal iff they are parallel (i.e. πππ β π‘ β
π π π’ππ π‘πππ‘ π₯ = π‘π¦).
Proof: (=>)
Let π₯ be π₯1 ,π₯2 and π¦ be π¦1 ,π¦2
Given π₯,π¦ are null vectors ,
π. π. , π₯12 + 2π΅π₯1π₯2 β π΄π₯2
2 = 0
π¦12 + 2π΅π¦1π¦2 β π΄π¦2
2 = 0
π₯1 + π΅π₯2 2 = (π΄ + π΅2)π₯2
2
π¦1 + π΅π¦2 2 = (π΄ + π΅2)π¦2
2
π₯1 + π΅π₯2 π¦1 + π΅π¦2 = (π΄ + π΅2)π₯2π¦2
πβ² π₯,π¦ = 0
Hence π₯,π¦ are parallel.
Proof :(<=)
Given π₯, π¦ are parallel
π. π. , π₯ = π‘π¦ πππ π πππ π‘ β π
πβ² π₯, π¦ = πβ² π‘π¦,π¦
= π‘πβ² π¦,π¦
= 0 (π ππππ π¦ ππ ππ’ππ π£πππ‘ππ)
Hence π₯,π¦ are orthogonal
Theorem 2: If π₯ β π₯π and πβ²(π₯ β π₯π) = 0 then
πΉππ,π = πΆπ(π₯) β© πΆπ(π₯0)
Proof:
Consider a element π¦ =π₯0 + π‘ π₯ β π₯0 ; of πΉππ,π
π¦- π₯0 = π‘ π₯ β π₯0
πβ²(π¦ -π₯0) = πβ²(π‘ π₯ β π₯0 )
= π‘2(πβ² π₯ β π₯0 )
= π‘2 0 ( given πβ² π₯ β π₯π = 0)
πβ²(π¦ -π₯0) = 0
So, π¦ belongs to πΆπ(π₯0)
10
Now let us prove πΉππ,π = πΉπ,ππ then we can say π¦ also belongs to πΆπ(π₯)
we know π π₯0,π₯ = { π₯0 + π‘(π₯ β π₯0); π‘ β π }
= { π‘π₯ β (π‘ β 1) π₯0); π‘ β π }
= { π₯ + (π‘ β 1)π₯ β (π‘ β 1) π₯0); π‘ β π }
= { π₯ + (π‘ β 1)(π₯ β π₯0)); π‘ β π }
= { π₯0 + π(π₯ β π₯0); π β π }
πΉππ,π = πΉπ,ππ
So, π¦ belongs to πΆπ(π₯)
Hence πΉππ,π β πΆπ(π₯) β© πΆπ(π₯0)
Now we have to prove πΆπ(π₯) β© πΆπ π₯0 β πΉππ,π
Consider a element π¦ β πΆπ(π₯) β© πΆπ π₯0 π¦ β π₯,π¦ β π₯0 πππ ππ’ππ π£πππ‘πππ
And we already now π₯ β π₯0 ππ πππ π π ππ’ππ π£πππ‘ππ (given πβ² π₯ β π₯0 = 0)
we know π¦ β π₯0 = π¦ β π₯ β π₯ β π₯0 πβ² π¦ β π₯0 = πβ²( π¦ β π₯ β π₯ β π₯0 )
= πβ²( π¦ β π₯ + π₯0 β π₯ )
πβ² π¦ β π₯0 = πβ² π¦ β π₯ + πβ² π₯0 β π₯ + 2π π¦ β π₯, π₯0 β π₯
π π¦ β π₯, π₯0 β π₯ = 0 (since Qβ π¦ β π₯0 = πβ² π¦ β π₯ = πβ² π₯0 β π₯ = 0)
π¦ β π₯, π₯0 β π₯ orthogonal
Case(i): π¦ = π₯ then π¦ β πΉππ,π
Case(ii): π¦ β π₯ then π¦ β π₯ must be parallel with π₯ β π₯0
i.e., π¦ β π₯ = π‘ π₯0 β π₯ π¦ = π₯ + π‘(π₯0 β π₯)
π¦ β πΉππ,π
So, πΆπ(π₯) β© πΆπ π₯0 β πΉππ,π
Hence πΉππ,π = πΆπ π₯ β© πΆπ π₯0 is proved
Theorem (3): Suppose that π£ = (π£1 ,π£2) is timelike and π€ = (π€1 ,π€2)is either timelike or null and
nonzero. Then either
(i) π£2π€2> π, in which case πβ²(π£,π€) < π,
or
(ii) π£2π€2< π, in which case πβ² π£,π€ > π.
11
Proof:
π£ ππ π‘πππππππ => π£12 + 2π΅ π£1π£2 < π΄π£2
2
Add π΅2π£22 on both the sides
π£12 + 2π΅ π£1π£2 + π΅2π£2
2 < π΄π£22 + π΅2π£2
2
π£1 + π΅π£2 2 < π΄π£2
2 + π΅2π£22 (π)
π€ ππ π‘πππππππππ ππ’ππ => π€12 + 2π΅ π€1π€2 β€ π΄π€2
2
Add π΅2π€22 on both the sides
π€12 + 2π΅ π€1π€2 + π΅2π€2
2 β€ π΄π€22 + π΅2π€2
2
π€1 + π΅π€2 2 β€ π΄π€2
2 + π΅2π€22 (ππ)
Multiplying (i) and (ii)
π£1 + π΅π£2 π€1 + π΅π€2 2 < π΄ + π΅2 2π£2
2π€22
π£1 + π΅π£2 π€1 + π΅π€2 < π΄ + π΅2 π£2π€2 (πππ)
if π£2π€2 > 0 by equation (iii) & π£1 + π΅π£2 π€1 + π΅π€2 β€ π£1 + π΅π£2 π€1 + π΅π€2
π£1 + π΅π£2 π€1 + π΅π€2 < π΄ + π΅2 π£2π€2
π£1π€1 + π΅ π£1π€2 + π£2π€1 + π΅2π£2π€2 < π΄π£2π€2 + π΅2π£2π€2
π£1π€1 + π΅ π£1π€2 + π£2π€1 β π΄π£2π€2 < 0
πβ² π£,π€ < 0
If π£2π€2 < 0 by equ (iii) & β π£1 + π΅π£2 π€1 + π΅π€2 β€ π£1 + π΅π£2 π€1 + π΅π€2
β π£1 + π΅π£2 π€1 + π΅π€2 < π΄ + π΅2 π£2π€2
π£1 + π΅π£2 π€1 + π΅π€2 > π΄ + π΅2 π£2π€2
π£1π€1 + π΅ π£1π€2 + π£2π€1 + π΅2π£2π€2 > π΄π£2π€2 + π΅2π£2π€2
π£1π€1 + π΅ π£1π€2 + π£2π€1 β π΄π£2π€2 > 0
πβ² π£,π€ > 0
Hence Proved.
Corollary: A nonzero vector in π 2 is πβ²-orthogonal to a timelike vector, then it must be spacelike.
Let us define a relation ~ on π as follows, ππ π£,π€ β π , then v~π w iff πβ²(π£,π€) < π .
Proof: It is directly implied from above theorem
Theorem (4): The relation, ~ is an equivalence on π, and π is the union of two disjoint sets denoted by
π+and πβ, which are cones.
Proof:
Reflexive: πβ² π£ < 0 πππ πππ π£ β π
π£ ~ π£ β π£ β π
Symmetric: π£ ~ π€ => πβ² π£,π€ < 0 => πβ² π€, π£ < 0 => π€ ~ π£
12
Transitive: let π£~π€ πππ π€~π§
πβ² π£,π€ < 0 πππ πβ² π€, π§ < 0
π£2π€2 > 0 πππ π€2π§2 > 0 (from theorem 3)
π£2π§2 > 0 ( for both π€2 > 0 πππ π€2 < 0 ππ‘ πππππ )
πβ² π£, π§ < 0
π£ ~ π§
Hence β²~β² is equivalence relation on π
π+ = {π£ = (π£1 , π£2) |π£ β π πππ π£2 > 0}
πβ = {π£ = (π£1 , π£2) |π£ β π πππ π£2 < 0}
To prove π+ and πβ are cones we have to show that π£,π€ β π+ (πβ) then for any positive real number π,
ππ£ , π£ + π€ β π+ (πβ)
We know ππ£2 > 0 => ππ£ β π+
Let us consider the case of π£ + π€, we know π£1 + π΅π£2 2 < π΄ + π΅2 π£2
2 & π€1 + π΅π€2 2 β€ (π΄ + π΅2)π€2
2
π£1 + π΅π£2 π€1 + π΅π€2 < π΄ + π΅2 π£2π€2
π£1π€1 + π΅ π£1π€2 + π£2π€1 + π΅2π£2π€2 < π΄π£2π€2 + π΅2π£2π€2
π£1π€1 + π΅ π£1π€2 + π£2π€1 β π΄π£2π€2 < 0
πβ² π£,π€ < 0 & π£ + π€ β π+
Hence π+, πβ are both disjoint cones
Theorem (5): (Reversed Schwartz Inequality) If π£ and π€ are timelike vectors in π 2 then
πβ² π£,π€ 2 β₯ [πβ²(π£, π£)πβ²(π€,π€)] ,
and equality holds if and only if π£ and π€ are linearly dependent.
Proof: Construct a vector π’
π’ = ππ£ β ππ€ where π = π π£,π€ and π = π(π£, π£)
consider π π’, π£ = π ππ£ β ππ€, π£
= ππ π£, π£ β ππ(π€, π£)
= π π£,π€ π π£, π£ β π(π£, π£)π(π€, π£)
= 0
π’, π£ are orthogonal
13
π’ is spacelike or zero vector (from theorem 3)
π π’,π’ β₯ 0
π ππ£ β ππ€,ππ£ β ππ€ > 0
π2π π£, π£ + π2π π€,π€ β 2πππ π£,π€ β₯ 0
π π£,π€ 2π π£, π£ + π π£, π£ 2π π€,π€ β₯ 2 π π£,π€ 2π π£, π£)
2π π£,π€ 2 β₯ π π£,π€ 2 + π π£, π£ π π€,π€ (cancelling π π£, π£ and since π π£, π£) < 0
π π£,π€ 2 β₯ π π£, π£ π π€,π€
Equality holds when π’ = 0
ππ£ β ππ€ = 0
ππ£ = ππ€
π£,π€ are linear
Hence proved
Theorem (6): (Reversed Triangle Inequality) Let π£ and π€ be timelike with πβ²(π£,π€) < π. Then
πβ (π£ + π€) β₯ πβ(π£) + πβ(π€),
and equality holds if and only if π£ and π€ are linearly dependent.
Proof: By theorem 5 πβ² π£,π€ 2 β₯ πβ² π£, π£ πβ² π€,π€ β₯ [βπβ²(π£, π£)][βπβ²(π€,π€)]
So, πβ² π£,π€ β₯ βπβ²(π£, π£) βπβ²(π€,π€)
πβ² π£,π€ β€ β βπβ²(π£, π£) βπβ²(π€,π€) (given πβ²(π£,π€) < π)
πβ² π£,π€ β€ β βπβ²(π£) βπβ²(π€) (i)
πβ π£ + π€ = βπβ²(π£ + π€)
πβ π£ + π€ = β(πβ² π£ + πβ² π£ + 2πβ² π£,π€ )
πβ π£ + π€ β₯ βπβ² π£ 2
+ βπβ² π€ 2
+ 2 βπβ²(π£) βπβ²(π€) (ππππ π )
πβ π£ + π€ β₯ πβ π£ 2+ πβ π€ 2 + 2πβ π£ πβ π€
πβ π£ + π€ β₯ πβ(π£) + πβ(π€)
Since we used Reversed Schwartz Inequality theorem in which equality holds only when π£ ,π€ are linear
itβs the same in this result also.
Lemma: Let πΌ: π΄,π΅ β π be a smooth timelike and future directed curve. Then for any π‘0 in π΄,π΅
there exist π > 0 such that (π‘0 β π , π‘0 + π ) is contained in π΄,π΅ , πΌ(π‘) is in past cone at πΌ(π‘0) for every
π‘ in (π‘0 β π , π‘0) and is in future cone at for every π‘ in (π‘0 , π‘0 + π ).
14
Proof: We prove that there exists a π1 > 0 such that πΌ π‘ is in πΆπ+(πΌ π‘0 ) for each π‘ in (π‘0 , π‘0 + π1 ). The
argument to produce and π2 > 0 with πΌ π‘ is in πΆπβ(πΌ π‘0 ) for each π‘ in (π‘0 β π2 , π‘0 ) is similar. Taking π
be the smallest of π1 and π2 proves the lemma
If no such π1 exists then one can produce a sequence π‘1 > π‘2 > π‘3 β¦ > π‘0 in π‘0 ,π΅ such that
limπββ π‘π = π‘0 and such that one of the following is true
1) πβ²(πΌ π‘π β πΌ π‘0 ) β₯ 0 for all π (i.e., πΌ π‘π β πΌ π‘0 is spacelike )
2) πβ²(πΌ π‘π β πΌ π‘0 ) < 0 but πΌ π‘π β πΌ π‘0 is past directed for every π
Consider case(1), On contrary let us assume that such a sequence doesnβt exist,
i.e., πβ² πΌ π‘π βπΌ π‘0
π‘πβ π‘0 β₯ 0 for all π
πβ² π₯1 π‘π βπ₯
1 π‘0
π‘πβ π‘0 ,π₯2 π‘π βπ₯
2 π‘0
π‘πβ π‘0 β₯ 0
Thus,
limπββ πβ² π₯1 π‘π βπ₯
1 π‘0
π‘πβ π‘0 ,π₯2 π‘π βπ₯
2 π‘0
π‘πβ π‘0 β₯ 0
πβ² ππ₯1 π‘0
ππ‘ ,ππ₯2 π‘0
ππ‘ β₯ 0
πβ² πΌβ² π‘0 β₯ 0
βπΌβ is spacelike or null which is contrary to the fact βπΌβ is timelike.
Now consider Case(2) :
If πβ²(πΌ π‘π β πΌ π‘0 ) < 0 but πΌ π‘π β πΌ π‘0 is past directed
by definition of πΌβ² π‘0 is future directed and from theorem(3) for all π
πβ²(πΌ π‘π β πΌ π‘0 , πΌβ² π‘0 ) > 0
limπββ
πβ² πΌ π‘π β πΌ π‘0
π‘π β π‘0 ,πΌβ² π‘0 > 0
πβ² limπββ
πΌ π‘π β πΌ π‘0
π‘π β π‘0 ,πΌβ² π‘0 > 0
πβ² πΌβ² π‘0 ,πΌβ² π‘0 > 0
πβ² πΌβ² π‘0 > 0
Which is again contrary to the fact πΌ is timelike. Hence the lemma is true.
15
Theorem (7): let π and π be in πΉπ then π β π is timelike and future directed if and only if there exists a
smooth future directed timelike curve πΌ: π, π β πΉπ such that πΌ π = π and πΌ π = π.
Proof: The necessity is clear since always there exists a straight line(curve) joining π and π. i.e., there
exists a curve πΌ: π, π β πΉπ such that πΌ π‘ = π π‘βπ
πβπ + π
(πβπ‘)
(πβπ)
To prove sufficiency we take πΌ as smooth future directed timelike extension of πΌ to some interval π΄,π΅
containing [a,b]. By the above lemma there exists an π1 > 0, such that
π,π + π1 β π΄,π΅ & πΌ π‘ β πΆπ+ πΌ π‘0 β π‘ β π,π + π1
Let π‘0 be supremum of all such π1. Since π < π΅ it will be sufficient to show π‘0 = π΅ for this let us assume
the contrary that π΄ < π‘0 < π΅. According to the above lemma,
βπ > 0 such that π‘0 β π, π‘0 + π β π΄,π΅ for π‘ ππ π‘0 β π , π‘0 and πΌ π‘ β πΆπβ πΌ π‘0 and for π‘ in
π‘0 , π‘0 + π and πΌ π‘ β πΆπ+ πΌ π‘0 . Now if πΌ π‘0 is in πΆπ
+ π then for any π‘ in π‘0 , π‘0 + π which implies
πΌ π‘0 β π + πΌ π‘ β πΌ π‘0 = πΌ π‘ β π
Which is timelike and future directed contradicting the definition of π‘0 . And if πΌ π‘0 is outside Null
cone at β²πβ² then π‘0 β π, π‘0 is disjoint from future cone at β²πβ² , which impliesπ‘0 = π΅ which proves the
theorem.
Definition: If πΌ: π° β πΉπ , the proper time function on π° = π, π is defined by
πβ π‘ = [πβ²(πΌβ² π’ ,πΌβ² π’ |]1/2 ππ’π
π
, π‘ β π°
For π₯, π₯0 β πΉπ with πβ² π₯ β π₯0 > 0, the proper spatial separation defined by πβ π₯ β π₯0 =
πβ² π₯ β π₯0 .
Theorem (8) : Let π₯, π₯0 , π₯1 β πΉπ for which π₯1 β π₯0, π₯1 β π₯ are spacelike and π₯ β π₯0 , π₯1 β π₯ are
orthogonal then ,
πβ2 π₯1 β π₯0 = πβ
2 π₯1 β π₯ β π2 π₯ β π₯0 .
Proof: π₯1 β π₯0 can be written as π₯1 β π₯ + π₯ β π₯0 . So,
πβ2 π₯1 β π₯0 = πβ
2 π₯1 β π₯ + π₯ β π₯0
= πβ² π₯ β π₯0 + πβ² π₯1 β π₯ β 2πβ²(π₯1 β π₯, π₯ β π₯0)
As π₯ β π₯0 , π₯1 β π₯ are orthogonal
16
= πβ² π₯ β π₯0 + πβ² π₯1 β π₯
= πβ² π₯ β π₯0 β (βπβ² π₯1 β π₯ )
= πβ2 π₯1 β π₯ + π₯ β π₯0
6.Topologies induced on πΉπ by g
We can consider the Topologies generated by the following Bases of topology
π©β = π·β π₯; π ; π₯ β πΉπ , π > 0 βͺ β & π©+ = π·+ π₯; π ; π₯ β πΉπ , π > 0 βͺ β
Where
π·β π₯; π = π¦ β πΉπ ;ππΈ π₯,π¦ < π πππ π π¦ β π₯,π¦ β π₯ < 0 βͺ π₯
π·+ π₯; π = π¦ β πΉπ ;ππΈ π₯,π¦ < π πππ π π¦ β π₯,π¦ β π₯ > 0 βͺ π₯
Let us denote the topologies generated by π©β and π©+, with πβand π+, respectively.
Concerning π we have three possibilities:
(i) π΅2 + π΄ > 0; (ii) π΅2 + π΄ < 0;
(iii) π΅2 + π΄ = 0
Case (i). Firstly we have
Lemma: π©β and π©+ are bases for certain topologies
Proof: Let us consider for example B_ (the case for /3+ is similar). We have to check the following two
properties:
a) πΉπ = π΅π© βπ©β ;
b) β π©1,π©2 β π©β,π©1 β© π©2 can be written as union of elements of π©β.
For a) it is obvious since for every π₯ β πΉπ β π·β π₯; π β π©β such that π₯ β π·β π₯; π
In order to prove b) let π©1 = π·β π₯1; π1 πππ π©2 = π·β π₯2; π2 . If π₯1 = π₯2 then (b) is obvious and the
case of π₯1 β π₯2 π©1 β©π©2 = β is also trivial.
So let us consider the case π©1 β©π©2 β β . It is easy to see that π©1 = ππ₯1 βͺ {π₯1 } , π©2 = ππ₯2
βͺ {π₯2 }
Where ππ₯π ={ π₯ β πΉπ ;ππΈ π₯, π₯π < ππ} β© {π₯ β πΉπ ;πβ² π₯ β π₯π ,π₯ β π₯π < 0} for π = 1,2 which are
obviously open subsets in the Euclidean topology on πΉπ. It follows,
π©1 β© π©2 = ππ₯1 β© ππ₯2
βͺ {π₯1} β© ππ₯2 βͺ ππ₯1
β© {π₯2}
Let π₯ β ππ₯1 β© ππ₯2
is a non empty openset, there exists ππ₯ > 0 such that
πΆ = { π₯ β πΉπ ;ππΈ π₯,π¦ < ππ₯} β ππ₯1 β© ππ₯2
Denoting
17
π΄ = { π¦ β πΉπ ;πβ² π₯ β π₯π ,π₯ β π₯π <0}
And
π·β π₯; ππ₯ = π΄ β© πΆ βͺ π₯ β ππ₯1 β© ππ₯2
It easily follows
ππ₯1 β© ππ₯2
= π·β π₯; ππ₯
π₯βππ₯1 β©ππ₯2
Now if π₯1 β ππ₯2 πππ π₯2 β ππ₯1
then
π©1 β© π©2 = ππ₯1 β© ππ₯2
= π·β π₯; ππ₯
π₯βππ₯1 β©ππ₯2
If for example π₯1 β ππ₯2 and π₯2 β ππ₯1
then because ππ₯1 is open there exists π1 < π1sufficiently small such
that
πΆ1 = {π₯ β πΉπ ;ππΈ π₯,π¦ < ππ₯} β ππ₯2
This immediately implies
π·β π₯1; π1 = πΆ1 β© {π₯ β πΉπ ;πβ² π₯βπ₯π ,π₯βπ₯π < 0} β ππ₯2
And
π·β π₯1; π1 \{π₯1} β ππ₯1
That is π·β π₯1; π1 \{π₯1} β ππ₯1 β© ππ₯2
. Consequently,
π©1 β© π©2 = ππ₯1 β© ππ₯2
βͺ {π₯1} = π·β π₯; ππ₯
π₯βππ₯1 β©ππ₯2
βͺ π·β π₯1; π1
The cases π₯1 β ππ₯2 and π₯2 β ππ₯1
and π₯1 β ππ₯2 and π₯2 β ππ₯1
are similar hence π©β is base
Similar argument can be extended to prove π©+ is base of certain topology which proves lemma.
Lemma: The closure of ππ of πΞ΅P (π₯) is πΆππΈ πΞ΅
P π₯ β (πππ¦πΈ πΞ΅P π₯ β© πππ¦πΈ π₯ )
Proof: Since π-Topology is finer than πΈ-Topology πΆππ π΄ β πΆππΈ π΄ for any set π΄ ππ π. Moreover the
points in πππ¦πΈ πΞ΅P π₯ β© πππ¦πΈ π₯ are not in πΆππ πΞ΅
P π₯ since if βπ¦β is such a point then πΞ΅/2P π¦
doesnβt intersect πΞ΅P π¦ (the null cone at βπ¦β is tangent to the surface of the Euclidean ball πΞ΅
E π₯
Thus, πΆππ πΞ΅P π₯ β πΆππΈ πΞ΅
P π₯ β (πππ¦πΈ πΞ΅P π₯ β© πππ¦πΈ π₯ ). Reverse containment is clear since if
βπ¦β is in the set on right hand side, every πΞ΄P π¦ intersects πΞ΅
P π₯ .
Theorem (9): (a) If B2 + A > 0 then the topologies πβand π+ generated by the bases π©β πππ π©+
respectively, are finer than the Euclidean topology, are Hausdorff, are connected, but not regular, and not
locally compact.
Proof:
i) πβand π+ are finer: Let ππΈbe Euclidean topology with base
π©πΈ = {π·πΈ π₯; π ; π₯ β πΉπ , π > 0 }
π€ππππ π·πΈ π₯; π = {π¦ β πΉπ ;ππΈ π₯,π¦ < π}
Take any element of base π©πΈ say π·πΈ π¦; ππ¦ . For any π₯ β π·πΈ π¦; ππ¦ β π©1 β π©β such that
π = π·β π₯; π where 0 < π < ππ¦ β ππΈ π¦; π₯ . Hence any π·πΈ π¦; ππ¦ is also open in πβ.
Conversely take any open set in base π©β of πβ say π·β π₯; π . Clearly there doesnβt exist any open
set in π©πΈ which contains βπ₯β and is contained in π·β π₯; π .
18
Hence πβ is finer topology than ππΈ . With the similar argument we can say π+ is finer than ππΈ
which proves the theorem.
ii) Consider Topology πβ any two points π₯ = (π₯1 ,π₯2 ) and π¦ = (π¦1 ,π¦2 )
Let π = ππΈ π₯,π¦ , take ππ₯ , ππ¦ < π/3 π·β π₯; ππ₯ and π·β π¦; ππ¦ are disjoint open sets in πβ .
Hence (πΉπ ,πβ) is hausdorff. Similarly (πΉπ ,π+) is also Hausdorff.
iii) Claim:πΉπ πππ β are the only two open and closed sets in (πΉπ ,πβ).
Take any open set in (say A). we can write
π΄ = π·β π₯π ; ππ
πβπ
π΄ = [ π¦ β πΉπ ;ππΈ π₯π ,π¦ < ππ πππ π π¦ β π₯π ,π¦ β π₯π < 0 βͺ
πβπ
{π₯π}]
And
π΄ = [ π¦ β πΉπ ;ππΈ π₯π ,π¦ β€ ππ πππ π π¦ β π₯π ,π¦ β π₯π β€ 0
πβπ
]
β π΄
Hence (πΉπ ,πβ) is connected. Similarly (πΉπ ,π+) is also connected.
iv) From the lemma from before it can be said that πΆππ πΞ΄P π₯ β πΞ΅
P π₯ for any value of Ξ΅. Hence
(πΉπ ,πβ) is not regular. Similarly, (πΉπ ,π+) is also not regular.
v) Since π- topology is finer than πΈ- topology, any π- compact is πΈ- compact. And no πΆππ πΞ΅P π₯
is compact (since not even πΈ- closed) we find that no point in (πΉπ ,πβ) has a compact
neighborhood. In particular (πΉπ ,πβ) is also not locally compact. Similarly, (πΉπ ,π+) is also not
locally compact.
7.Future scope:
My project is to understand the works of Naber and Paul thoroughly and work upon the
challenges in this explorable field of mathematics. The above study comprised of Geometrical properties
Topologies induced by the inner productβ πβ on different 2 dimensional systems. And similarly we can
study other characteristics like trigonometry and physics generated by these complex type numbers which
are extensively used in Relativity and extend the results in this study to the higher dimensional vector
spaces.
References: 1. Einstein Albert, "SpaceβTime", Encyclopedia Britannica (1926), 13th ed
2. Feynman, Richard Phillips; MorΓnigo, Fernando B.; Wagner, William; Pines, David; Hatfield, Brian
(2002). βFeynman Lectures on Gravitationβ. West view Press.
3. Naber G.L., βThe Geometry of Minskowski Spacetime . An Introduction to Mathematics of
Special Theory of Relativity β Springer- verlag, Newyork (1992)
4. Paul Flejstad, βTwo-Dimensional Geometries, Topologies, Trigonometry and Physics generated
by Complex-Type Numbersβ (2001).
5. Stephen Willard, βGeneral Topologyβ, Addison Wisley (1970)