Parallel Bundles in Planar Map Geometries
Parallel Bundles in Planar Map Geometries
Parallel Bundles in Planar Map Geometries
1. Introduction
A map is a connected topological graph cellularly embedded in a surface. On the past century, many works are concentrated on to nd the combinatorial properties of maps, such as to determine whether exists a particularly embedding on a surface ([7], [11]) or to enumerate a family of maps ([6]). All these works are on the side of algebra, not the object itself, i.e., geometry. For the later, more attentions are given to its elements behaviors, such as, the line, angle, area, curvature, , see also [12] and [14]. For returning to its original face, the conception of map geometries is introduced in [10]. It is proved in [10] that the map geometries are nice model of the Smarandache geometries. They are also a new kind of intrinsic geometry of surfaces ([1]). The main purpose of this paper is to determine the behaviors of parallel bundles in planar geometries, a generalization of parallel lines in the Euclid plane geometry. An axiom is said Smarandachely denied if the axiom behaves in at least two dierent ways within the same space, i.e., validated and invalided, or only invalided but in multiple distinct ways. A Smarandache geometry is a geometry which has at least one Smarandachely denied axiom(1969)([5], [13]). In [3] and [4], Iseri presented a nice model of the Smarandache geometries, called smanifolds by using equilateral triangles, which is dened as follows([3], [5] and [9]): An s-manifold is any collection C (T, n) of these equilateral triangular disks Ti , 1 i n satisfying the following conditions: (i) Each edge e is the identication of at most two edges ei , ej in two distinct triangular disks Ti , Tj , 1 i, j n and i = j ; (ii) Each vertex v is the identication of one vertex in each of ve, six or seven distinct triangular disks.
Vol. 1
121
The conception of map geometries without boundary is dened as follows ([10]). Denition 1.1 For a given combinatorial map M , associates a real number (u), 0 < (u) < , to each vertex u, u V (M ). Call (M, ) a map geometry without boundary, (u) the angle factor of the vertex u and to be orientablle or non-orientable if M is orientable or not. In [10], it has proved that map geometries are the Smarandache geometries. The realization of each vertex u, u V (M ) in R3 space is shown in the Fig.1 for each case of (u)(u) > 2 , = 2 or < 2 , call elliptic point, euclidean point and hyperbolic point, respectively.
(u)(u) = 2
(u)(u) > 2
Therefore, a line passes through an elliptic vertex, an euclidean vertex or a hyperbolic vertex (u) u has angle (u)2 at the vertex u. It is not 180 if the vertex u is elliptic or hyperbolic. Then what is the angle of a line passes through a point on an edge of a map? It is 180 ? Since we wish the change of angles on an edge is smooth, the answer is not. For the Smarandache geometries, the parallel lines in them are need to be given more attention. We have the following denition. Denition 1.2 A family L of innite lines not intersecting each other in a planar geometry is called a parallel bundle. In the Fig.2, we present all cases of parallel bundles passing through an edge in planar geometries, where, (a) is the case of points u, v are same type with (u)(u) = (v )(v ), (b) and (c) the cases of same types with (u)(u) > (v )(v ) and (d) the case of u is elliptic and v hyperbolic.
Fig. 2 Here, we assume the angle at the intersection point is in clockwise, that is, a line passing
122
Linfan Mao
No. 2
through an elliptic point will bend up and a hyperbolic point will bend down, such as the cases (b),(c) in the Fig.2. For a vector O on the Euclid plane, call it an orientation. We classify parallel bundles in planar map geometries along an orientation O .
0.
+
Proof. If L is a parallel bundle, then any two lines L1 , L2 will not intersect after them passing through the edge uv . Therefore, if 1 , 2 are the angles of L1 , L2 at the intersect points of L1 , L2 with uv and L2 is far from u than L2 , then we know that 2 1 . Whence, for any point with x distance from u and x > 0, we have that
=
+
meeting uv , whence, any two lines in L will not intersect each other after them passing through uv . Therefore, L is a parallel bundle. A general condition for a family of parallel lines passing through a cut of a planar map being a parallel bundle is the following. Proposition 2.2 Let (M, ) be a planar map geometry, C = {u1 v1 , u2 v2 , , ul vl } a cut of the map M with order u1 v1 , u2 v2 , , ul vl from the left to the right, l 1 and the angle functions on them are f1 , f2 , , fl , respectively, also see the Fig.3.
Vol. 1
123
Fig. 3 Then a family L of parallel lines passing through C is a parallel bundle i for any x, x 0,
f1 (x) 0 f1+ (x) + f2+ (x) 0 f1+ (x) + f2+ (x) + f3+ (x) 0 f1+ (x) + f2+ (x) + + fl+ (x) 0. Proof. According to the Proposition 2.1, see the following Fig.4,
Fig. 4 we know that any lines will not intersect after them passing through u1 v1 and u2 v2 i for x > 0 and x 0, f2 (x + x) + f1+ (x)x f2 (x). That is, f1+ (x) + f2+ (x) 0. Similarly, any lines will not intersect after them passing through u1 v1 , u2 v2 and u3 v3 i for x > 0 and x 0, f3 (x + x) + f2+ (x)x + f1+ (x)x f3 (x).
Linfan Mao
No. 2
f1+ (x) + f2+ (x) + f3+ (x) 0. Generally, any lines will not intersect after them passing through u1 v1 , u2 v2 , , ul1 vl1 and ul vl i for x > 0 and x 0, fl (x + x) + fl1+ (x)x + + f1+ (x)x fl (x). Whence, we get that f1+ (x) + f2+ (x) + + fl+ (x) 0. Therefore, a family L of parallel lines passing through C is a parallel bundle i for any x, x 0, we have that
f1 (x) 0 f1+ (x) + f2+ (x) 0 f1+ (x) + f2+ (x) + f3+ (x) 0 f1+ (x) + f2+ (x) + + fl+ (x) 0. This completes the proof. . Corollary 2.1 Let (M, ) be a planar map geometry, C = {u1 v1 , u2 v2 , , ul vl } a cut of the map M with order u1 v1 , u2 v2 , , ul vl from the left to the right, l 1 and the angle functions on them are f1 , f2 , , fl . Then a family L of parallel lines passing through C is still parallel lines after them leaving C i for any x, x 0,
f1 (x) 0 f1+ (x) + f2+ (x) 0 f1+ (x) + f2+ (x) + f3+ (x) 0 f1+ (x) + f2+ (x) + + fl1+ (x) 0 f1+ (x) + f2+ (x) + + fl+ (x) = 0. Proof. According to the Proposition 2.2, we know the condition is a necessary and sucient condition for L being a parallel bundle. Now since lines in L are parallel lines after them leaving C i for any x 0 and x 0, there must be that fl (x + x) + fl1+ (x)x + + f1+ (x)x = fl (x). Therefore, we get that
Vol. 1
125
f1+ (x) + f2+ (x) + + fl+ (x) = 0 When do the parallel lines parallel the initial parallel lines after them passing through a cut C in a planar map geometry? The answer is in the following result. Proposition 2.3 Let (M, ) be a planar map geometry, C = {u1 v1 , u2 v2 , , ul vl } a cut of the map M with order u1 v1 , u2 v2 , , ul vl from the left to the right, l 1 and the angle functions on them are f1 , f2 , , fl . Then the parallel lines parallel the initial parallel lines after them passing through C i for x 0,
f1 (x) 0 f1+ (x) + f2+ (x) 0 f1+ (x) + f2+ (x) + f3+ (x) 0 f1+ (x) + f2+ (x) + + fl1+ (x) 0 and f1 (x) + f2 (x) + + fl (x) = l. Proof. According to the Proposition 2.2 and Corollary 2.1, we know the parallel lines passing through C is a parallel bundle. We calculate the angle (i, x) of a line L passing through an edge ui vi , 1 i l with the line before it meeting C at the intersection of L with the edge ui vi , where x is the distance of the intersection point to u1 on u1 v1 , see also the Fig.4. By the denition, we know the angle (1, x) = f (x) and (2, x) = f2 (x) ( f1 (x)) = f1 (x) + f2 (x) . Now if (i, x) = f1 (x) + f2 (x) + + fi (x) (i 1) , then similar to the case i = 2, we know that (i + 1, x) = fi+1 (x) ( (i, x)) = fi+1 (x) + (i, x) . Whence, we get that (i + 1, x) = f1 (x) + f2 (x) + + fi+1 (x) i. Notice that a line L parallel the initial parallel line after it passing through C i (l, x) = , i.e., f1 (x) + f2 (x) + + fl (x) = l. This completes the proof.
126
Linfan Mao
No. 2
Proposition 3.1 The angle function fl (x) of a line L passing through an edge uv at the point with distance x to u is fl (x) = (1 x (u)(v ) x (v )(v ) ) + , d(uv ) 2 d(uv ) 2
where, d(uv ) is the length of the edge uv . Proof. Since fl (x) is linear, we know that fl (x) satises the following equation.
(u)(u) 2 (u)(u) (v )(v ) 2 2
fl (x)
x , d(uv )
Corollary 3.1 Under the linear assumption, a family L of parallel lines passing through an edge uv is a parallel bundle i (v ) (u) . (v ) (u) Proof. According to the Proposition 2.1, a family of parallel lines passing through an edge uv is a parallel bundle i for x, x 0, f (x) 0, i.e., (v )(v ) (u)(u) 0. 2d(uv ) 2d(uv ) Therefore, a family L of parallel lines passing through an edge uv is a parallel bundle i (v )(v ) (u)(u). Whence, (u) (v ) . (v ) (u) For a family of parallel lines pass through a cut, we have the following condition for it being a parallel bundle. Proposition 3.2 Let (M, ) be a planar map geometry, C = {u1 v1 , u2 v2 , , ul vl } a cut of the map M with order u1 v1 , u2 v2 , , ul vl from the left to the right, l 1. Then under the linear assumption, a family L of parallel lines passing through C is a parallel bundle i the angle factor satises the following linear inequality system (v1 )(v1 ) (u1 )(u1 ) (v1 )(v1 ) (v2 )(v2 ) (u1 )(u1 ) (u2 )(u2 ) + + d(u1 v1 ) d(u2 v2 ) d(u1 v1 ) d(u2 v2 )
Vol. 1
127
(v2 )(v2 ) (vl )(vl ) + + d(u2 v2 ) d(ul vl ) (u1 )(u1 ) (u2 )(u2 ) (ul )(ul ) + + + . d(u1 , v1 ) d(u2 , v2 ) d(ul , vl )
Proof. Under the linear assumption, for any integer i, i 1, we know that fi+ (x) = (vi )(vi ) (ui )(ui ) 2d(ui vi )
by the Proposition 3.1. Whence, according to the Proposition 2.2, we get that a family L of parallel lines passing through C is a parallel bundle i the angle factor satises the following linear inequality system (v1 )(v1 ) (u1 )(u1 ) (v1 )(v1 ) (v2 )(v2 ) (u1 )(u1 ) (u2 )(u2 ) + + d(u1 v1 ) d(u2 v2 ) d(u1 v1 ) d(u2 v2 )
(v2 )(v2 ) (vl )(vl ) + + d(u2 v2 ) d(ul vl ) (u1 )(u1 ) (u2 )(u2 ) (ul )(ul ) + + + . d(u1 , v1 ) d(u2 , v2 ) d(ul , vl )
This completes the proof. For planar maps underlying a regular graph, we have the following interesting results for parallel bundles. Corollary 3.2 Let (M, ) be a planar map geometry with M underlying a regular graph, C = {u1 v1 , u2 v2 , , ul vl } a cut of the map M with order u1 v1 , u2 v2 , , ul vl from the left to the right, l 1. Then under the linear assumption, a family L of parallel lines passing through C is a parallel bundle i the angle factor satises the following linear inequality system (v1 ) (u1 ) (v1 ) (v2 ) (u1 ) (u2 ) + + d(u1 v1 ) d(u2 v2 ) d(u1 v1 ) d(u2 v2 ) (v1 ) (v2 ) (vl ) (u1 ) (u2 ) (ul ) + + + + + + d(u1 v1 ) d(u2 v2 ) d(ul vl ) d(u1 v1 ) d(u2 v2 ) d(ul vl ) and particularly, if assume that all the lengths of edges in C are the same, then
128
Linfan Mao
No. 2
Certainly, by choosing dierent angle factors, we can also get combinatorial conditions for existing parallel bundles under the linear assumption. Proposition 3.3 Let (M, ) be a planar map geometry, C = {u1 v1 , u2 v2 , , ul vl } a cut of the map M with order u1 v1 , u2 v2 , , ul vl from the left to the right, l 1. If for any integer i, i 1, (vi ) (ui ) , (vi ) (ui ) then under the linear assumption, a family L of parallel lines passing through C is a parallel bundle. Proof. Notice that under the linear assumption, for any integer i, i 1, we know that fi+ (x) = (vi )(vi ) (ui )(ui ) 2d(ui vi )
by the Proposition 3.1. Whence, fi+ (x) 0 for i = 1, 2, , l. Therefore, we get that
f1 (x) 0 f1+ (x) + f2+ (x) 0 f1+ (x) + f2+ (x) + f3+ (x) 0 f1+ (x) + f2+ (x) + + fl+ (x) 0. By the Proposition 2.2, we know that a family L of parallel lines passing through C is a parallel bundle.
Vol. 1
129
(2) C0100 : for any cut C along O , f (C ) < |C | ; (3) C0010 : for any cut C along O , f (C ) > |C | ; (4) C0001 : for any cut C along O , f+ (C, x) > 0 for x, x 0; (5) C1100 : There exist cuts C1 , C2 along O , such that f (C1 ) = |C1 | and f (C2 ) = c < |C2 | ; (6) C1010 : there exist cuts C1 , C2 along O , such that f (C1 ) = |C1 | and f (C2 ) > |C2 | ; (7) C1001 : there exist cuts C1 , C2 along O , such that f (C1 ) = |C1 | and f+ (C2 , x) > 0 for x, x 0; (8) C0110 : there exist cuts C1 , C2 along O , such that f (C1 ) < |C1 | and f (C2 ) > |C2 | ; (9) C0101 : there exist cuts C1 , C2 along O , such that f (C1 ) < |C1 | and f+ (C2 , x) > 0 for x, x 0; (10) C0011 : there exist cuts C1 , C2 along O , such that f (C1 ) > |C1 | and f+ (C2 , x) > 0 for x, x 0; (11) C1110 : there exist cuts C1 , C2 and C3 along O , such that f (C1 ) = |C1 | , f (C2 ) < |C2 | and f (C3 ) > |C3 | ; (12) C1101 : there exist cuts C1 , C2 and C3 along O , such that f (C1 ) = |C1 | , f (C2 ) < |C2 | and f+ (C3 , x) > 0 for x, x 0; (13) C1011 : there exist cuts C1 , C2 and C3 along O , such that f (C1 ) = |C1 | , f (C2 ) > |C2 | and f+ (C1 , x) > 0 for x, x 0; (14) C0111 : there exist cuts C1 , C2 and C3 along O , such that f (C1 ) < |C1 | , f (C2 ) > |C2 | and f+ (C1 , x) > 0 for x, x 0; (15) C1111 : there exist cuts C1 , C2 , C3 and C4 along O , such that f (C1 ) = |C1 | , f (C2 ) < |C2 | , f (C3 ) > |C3 | and f+ (C4 , x) > 0 for x, x 0. Notice that only the rst three classes may be parallel lines after them passing through the cut C . All of the other classes are only parallel bundles, not parallel lines in the usual meaning. Proposition 4.1 For an orientation O , the 15 classes C1000 C1111 are all the parallel bundles in planar map geometries. Proof. Not loss of generality, we assume C1 , C2 , , Cm , m 1, are all the cuts along O in a planar map geometry (M, ) from the upon side of O to its down side. We nd their structural characters for each case in the following discussion. C1000 : By the Proposition 2.3, a family L of parallel lines parallel their initial lines before meeting M after the passing through M . C0100 : By the denition, a family L of parallel lines is a parallel bundle along O only if f (C1 ) f (C2 ) f (Cm ) < . Otherwise, some lines in L will intersect. According to the Corollary 2.1, they parallel each other after they passing through M only if f (C1 ) = f (C2 ) = = f (Cm ) < . C0010 : Similar to the case C0100 , a family L of parallel lines is a parallel bundle along O only if
130
Linfan Mao
No. 2
< f (C1 ) f (C2 ) f (Cm ) and parallel each other after they passing through M only if < f (C1 ) = f (C2 ) = = f (Cm ). C0001 : Notice that by the proof of the Proposition 2.3, a line has angle f (C, x) (|C | 1) after it passing through C with the initial line before meeting C . In this case, a family L of parallel lines is a parallel bundle along O only if for xi , xi 0, 1 i m, f (C1 , x1 ) f (C2 , x2 ) f (Cm , xm ). Otherwise, they will intersect. C1100 : In this case, a family L of parallel lines is a parallel bundle along O only if there is an integer k, 2 k m, such that f (C1 ) f (C2 ) f (Ck1 ) < f (Ck ) = f (Ck+1 ) = = f (Cm ) = . Otherwise, they will intersect. C1010 : Similar to the case C1100 , in this case, a family L of parallel lines is a parallel bundle along O only if there is an integer k, 2 k m, such that = f (C1 ) = f (C2 ) = = f (Ck ) < f (Ck+1 ) f (Cm ). Otherwise, they will intersect. C1001 : In this case, a family L of parallel lines is a parallel bundle along O only if there is an integer k, l, 1 k < l m, such that for xi , xi 0, 1 i k or l i m,
f (C1 , x1 ) f (C2 , x2 ) f (Ck , xk ) < f (Ck+1 ) = f (Ck+2 ) = = f (Cl1 ) = < f (Cl , xl ) f (Cm , xm ). Otherwise, they will intersect. C0110 : In this case, a family L of parallel lines is a parallel bundle along O only if there is integers k, 1 k < m, such that f (C1 ) f (C2 ) f (Ck ) < < f (Ck+1 ) f (Cm ). Otherwise, they will intersect. C0101 : In this case, a family L of parallel lines is a parallel bundle along O only if there is integers k, 1 k m, such that for xi , xi 0, 1 i m,
f (C1 , x1 ) f (C2 , x2 ) f (Ck , xk ) < f (Ck+1 , xk+1 ) f (Cm , xm ), and there must be a constant in f (C1 , x1 ), f (C2 , x2 ), , f (Ck , xk ).
Vol. 1
131
C0011 : In this case, the situation is similar to the case C0101 and there must be a constant in f (Ck+1 , xk+1 ), f (Ck+2 , xk+2 ), , f (Cm , xm ). C1110 : In this case, a family L of parallel lines is a parallel bundle along O only if there is an integer k, l, 1 k < l m, such that
f (C1 ) f (C2 ) f (Ck ) < f (Ck+1 ) = = f (Cl1 ) = < f (Cl ) f (Cm ). Otherwise, they will intersect. C1101 : In this case, a family L of parallel lines is a parallel bundle along O only if there is an integer k, l, 1 k < l m, such that for xi , xi 0, 1 i k or l i m,
f (C1 , x1 ) f (C2 , x2 ) f (Ck , xk ) < f (Ck+1 ) = = f (Cl1 ) = < f (Cl , xl ) f (Cm , xm ) and there must be a constant in f (C1 , x1 ), f (C2 , x2 ), , f (Ck , xk ). Otherwise, they will intersect. C1011 : In this case, a family L of parallel lines is a parallel bundle along O only if there is an integer k, l, 1 k < l m, such that for xi , xi 0, 1 i k or l i m,
f (C1 , x1 ) f (C2 , x2 ) f (Ck , xk ) < f (Ck+1 ) = = f (Cl1 ) = < f (Cl , xl ) f (Cm , xm ) and there must be a constant in f (Cl , xl ), f (Cl+1 , xl+1 ), , f (Cm , xm ). Otherwise, they will intersect. C0111 : In this case, a family L of parallel lines is a parallel bundle along O only if there is an integer k, 1 k m, such that for xi , xi 0,
f (C1 , x1 ) f (C2 , x2 ) f (Ck , xk ) < < f (Cl , xl ) f (Cm , xm ) and there must be a constant in f (C1 , x1 ), f (C2 , x2 ), , f (Ck , xk ) and a constant in f (Cl , xl ), f (Cl+1 , xl+1 ), , f (Cm , xm ). Otherwise, they will intersect. C1111 : In this case, a family L of parallel lines is a parallel bundle along O only if there is an integer k, l, 1 k < l m, such that for xi , xi 0, 1 i k or l i m,
f (C1 , x1 ) f (C2 , x2 ) f (Ck , xk ) < f (Ck+1 ) = = f (Cl1 ) = < f (Cl , xl ) f (Cm , xm ) and there must be a constant in f (C1 , x1 ), f (C2 , x2 ), , f (Ck , xk ) and a constant in f (Cl , xl ), f (Cl+1 , xl+1 ), , f (Cm , xm ). Otherwise, they will intersect.
132
Linfan Mao
No. 2
Following the structural characters of the classes C1000 C1111 , by the Proposition 2.2, 2.3 and Proposition 3.1, we know that any parallel bundle is in one of the classes C1000 C1111 and each class in C1000 C1111 is non-empty. This completes the proof. A example of parallel bundle in a planar map geometry is shown in the Fig.5, in where the number on a vertex u denotes the number (u)(u).
Fig. 5
5. Generalization
All the planar map geometries considered in this paper are without boundary. For planar map geometries with boundary, i.e., some faces are deleted ([10]), which are correspondence with the maps with boundary ([2]). We know that they are the Smarandache non-geometries, satisfying one or more of the following conditions: (A1 )It is not always possible to draw a line from an arbitrary point to another arbitrary point. (A2 )It is not always possible to extend by continuity a nite line to an innite line. (A3 )It is not always possible to draw a circle from an arbitrary point and of an arbitrary interval. (A4 )Not all the right angles are congruent. (A5 )If a line, cutting two other lines, forms the interior angles of the same side of it strictly less than two right angle, then not always the two lines extended towards innite cut each other in the side where the angles are strictly less than two right angle. Notice that for an one face planar map geometry (M, )1 with boundary, if we choose all points being euclidean, then (M, )1 is just the Poincar es model for the hyperbolic geometry.
Vol. 1
133
Using the neutrosophic logic idea, we can also dene the conception of neutrosophic surface as follow, comparing also with the surfaces in [8] and [14]. Denition 5.1 A neutrosophic surface is a Hausdor, connected, topological space S such that every point v is elleptic, euclidean, or hyperbolic. For this kind of surface, we present the following problem for the further researching. Problem 5.1 To determine the behaviors of elements, such as, the line, angle, area, , in neutrosophic surfaces. Notice that results in this paper are just the behaviors of line bundles in a neutrosophic plane.
References
[1]A.D.Aleksandrov and V.A.Zalgaller, Intrinsic geometry of surfaces, American Mathematical Society, 1967. [2]R.P.Bryant and D.Singerman, Foundations of the theory of maps on surfaces with boundary, Quart.J.Math.Oxford, 36(2)(1985), 17-41. [3]H.Iseri, Smarandache manifolds, American Research Press, Rehoboth, NM, 2002. [4]H.Iseri, Partially Paradoxist Smarandache Geometries, http://www.gallup.unm. edu/ smarandache/Howard-Iseri-paper.htm. [5]L.Kuciuk and M.Antholy, An Introduction to Smarandache Geometries, Mathematics Magazine, Aurora, Canada,12(2003) [6]Y.P.Liu, Enumerative Theory of Maps, Kluwer Academic Publisher, Dordrecht, Boston, London, (1999). [7]Y.P.Liu, Embeddability in Graphs, Kluwer Academic Publisher, Dordrecht, Boston, London, (1995). [8]Mantredo P.de Carmao, Dierential Geometry of Curves and Surfaces, Pearson Education asia Ltd, (2004). [9]L.F.Mao, Automorphism groups of maps, surfaces and Smarandache geometries, American Research Press, Rehoboth, NM, 2005. [10]L.F.Mao, A new view of combinatorial maps by Smarandaches notion, arXiv, Math.GM/0506232, will also appear in Smarandache Notions Journal. [11]B.Mohar and C.Thomassen, Graphs on Surfaces, The Johns Hopkins University Press, London, 2001. [12]V.V.Nikulin and I.R.Shafarevlch, Geometries and Groups, Springer-Verlag Berlin Heidelberg, (1987). [13]F. Smarandache, Mixed noneuclidean geometries, eprint arXiv, math/0010119, 10/2000. [14]J.Stillwell, Classical topology and combinatorial group theory, Springer-Verlag New York Inc., (1980).