Introduction to Combinatorial and Computational Geometry

Introduction to Combinatorial and Computational Geometry 2022/2023

Zaklady kombinatoricke a vypocetni geometrie 2022/2023

Lecture: Wednesday at 9:00 in S4. The lectures are in English this year, but you can also ask questions in Czech or Slovak.
Prednaska: Ve stredu od 9:00 v S4. Prednaska je letos v anglictine, muzete se ale ptat i cesky nebo slovensky.

Exercise sessions: after the lecture at 10:40 in S9 according to the schedule on the exercise webpage.
Cviceni: po prednasce 10:40 v S9 podle rozvrhu na strance cviceni.

Extent of teaching: WINTER semester 2/2 (Exam and Exercise credit). Entry in SIS, syllabus
Rozsah vyuky: ZIMNI semestr 2/2 Z, Zk. Stranka v SISu, sylabus

Annotation
Discrete geometry investigates combinatorial properties of geometric objects such as finite point sets or convex sets in Euclidean spaces. Computational geometry considers the design of efficient algorithms for computing with geometric configurations, and discrete geometry serves as its mathematical foundation. Part I of the course is a concise introduction. The lecture will be very similar to the last year. The contents of Part II varies among the years, each year covering a few selected topics in more depth.

Anotace
Vypocetni geometrie se zabyva navrhem efektivnich algoritmu pro geometricke problemy v rovine i ve vicedimenzionalnim prostoru (napr. je-li dano n bodu v rovine, jak co nejefektivneji najit dvojici bodu s nejmensi vzdalenosti). Takove problemy jsou motivovany aplikacemi v pocitacove grafice, prostorovem modelovani (napr. molekul, budov, soucastek), geografickych informacnich systemech a pod. Pri analyze takovych algoritmu se potrebuje kombinatoricka geometrie, studujici kombinatoricke vlastnosti geometrickych konfiguraci, konvexnich mnozin a pod. Vysledky jsou dulezite i z ciste matematickeho hlediska, napr. v teorii cisel. V teto uvodni prednasce se probiraji zakladni pojmy a metody, s durazem na matematicky zaklad (jine mozne podani by bylo z vice "informatickeho" hlediska, s durazem na datove struktury, implementaci algoritmu apod.). O naplni prednasky si muzete udelat lepsi predstavu podle latky probirane v minulych letech.

Literature

[M1] Jiri Matousek's lecture notes Introduction to Discrete Geometry, ITI Series 2003-150, available in the Computer Science library of MFF UK. Here is a postscript file with two pages of text on one A4 page, intended for printing. Here is a PDF file for readers and tablets.
The lecture notes are, in fact, a selection of topics from the book J. Matousek: Lectures on Discrete Geometry.
A text for the part about incremental geometric algorithms is available here (12 pages).
Chan's algorithm for the construction of the convex hull: doi:10.1007/BF02712873
A standard introductory text about computational geometry: M. de Berg, M. van Kreveld, M. Overmars, and O. Schwarzkopf: Computational Geometry: Algorithms and Applications, Springer-Verlag, Berlin, 1997. The computer science library of MFF UK has a few copies.
J. Pach, P. Agarwal: Combinatorial Geometry, Cambridge University Press 1995

Literatura

Skripta Jiriho Matouska Introduction to Discrete Geometry vysla v ITI Seriich (preprintova rada Institutu teoreticke informatiky MFF UK) pod cislem 2003-150 a je k dispozici v informaticke knihovne MFF UK. Tady jsou take jako postscriptovy soubor, v nemz jsou pro usporu mista 2 male stranky na jedne strance A4. Je urcen hlavne pro dvoustranny tisk a ma vlevo okraj na svazani. A zde PDF soubor mysleny pro ctecky a tablety.
Predchozi text je ve skutecnosti vyber casti obvykle probiranych v prednasce z knihy J. Matousek: Lectures on Discrete Geometry.
Text k casti o inkrementalnich geometrickych algoritmech je k dispozici zde (12 str).
Chanuv algoritmus pro konstrukci konvexniho obalu: doi:10.1007/BF02712873
Kapitola "Geometricke algoritmy" ve skriptech Martina Marese
Standardni uvodni text o vypocetni geometrii je M. de Berg, M. van Kreveld, M. Overmars, and O. Schwarzkopf: Computational Geometry: Algorithms and Applications, Springer-Verlag, Berlin, 1997. Informaticke oddeleni knihovny ma nekolik exemplaru.
J. Pach, P. Agarwal: Combinatorial Geometry, Cambridge University Press 1995

Topics covered:

5.10. (PV)

Introductory information
Basic notions: d-dimensional Euclidean space R^d (the set of d-tuples of real numbers), hyperplane, closed halfspace
Afinne subspace, afinne hull, afinne combination, afinne (in)dependence
Convex set, convex hull, it is the set of all convex combinations
Caratheodory's theorem (only statement, the proof is an exercise)
Hyperplane separation theorem, sketch of proof of strict separation for a compact set vs. a closed set, and non-strict separation for arbitrary sets (to be continued next week)

12. 10. (PV), Literature: [M1, 1.3 + 1.4]

Rest of the proof of the hyperplane separation theorem
Afinne subspace, afinne hull, afinne combination, afinne (in)dependence
Radon's theorem
Helly's theorem
Infinite version of Helly's theorem
Centerpoint theorem (the proof will be given next week)

19. 10. (PV), Literature: [M1, 1.4 + 4.1 + 4.3]

Proof of the centerpoint theorem
The ham sandwich theorem (only statement)
Center transversal theorem (only statement)
Geometric graph, crossing number of a geometric graph
Crossing lemma (for geometric graphs)
Szemeredi–Trotter theorem about maximum number of incidences between points and lines

26.10. (PV), Literature: [M1, 4.1 + 4.2 + 2.1]

Proof of Szemeredi–Trotter theorem about maximum number of incidences between points and lines
Lower bound on the number of incidences between n points and n lines (idea of one of the constructions)
Upper bound O(n^{4/3}) on the number of unit distances (only statement & few words about the idea of the proof)
Lower bound n^{1+c/(log log n)} on the number of unit distances (short description of the construction, no proof)
Minkowski's theorem for the integer lattice

2.11 (PV), Literature: [M1, 2.2, 5.1, 5.2]

Application 1: looking out of a regular forest
Application 2: Diophantine approximation of real numbers
Lattice with basis z_1, z_2, ..., z_d, the determinant of a lattice
Minkowski's theorem for general lattices (only few words about the proof - it can be found in [M1])
An application of the general Minkowski's theorem in number theory: each prime number of the form 4k + 1 can be written as a sum of two squares (only the statement)
Geometric duality: duality transform D_0 between points (except 0) and hyperplanes avoiding the origin, duality lemma (duality preserves incidences), dual set X*

16.11. (JK)

Duality transform D between points and nonvertical hyperplanes
Convex polytopes: V-polytope, H-polyhedron, H-polytope, dimension of a polytope
Basic examples: d-dimensional cube, crosspolytope and simplex
Every H-polytope is a V-polytope
Every V-polytope is a H-polytope (sketch of a proof using duality and reverse implication)
Face of a polytope P, facet, edge, vertex
Extremal point of a set
The extremal points of P are exactly its vertices, P is the convex hull of its vertices (proof of two inclusions out of three)

23.11. (JK)

Proof of the third inclusion: if V is a minimal set such that P is a convex hull of P, then every point in V is a vertex of P.
The vertices of a face F of P are exactly those vertices of P that lie in F
The faces of a face F of P are exactly those faces of P that lie in F (sketch of proof)
Graph of a polytope, the graph of a 3-dimensional polytope is planar
Balinski's theorem (only statement)
Steinitz's theorem (only statement)
Face lattice of a polytope and its properties (without proof), combinatorial equivalence of polytopes
Corollary: the combinatorial type of a convex polytope is determined by the vertex-facet incidences (sketch of proof)
Dual polytope P*, the face lattice of a dual polytope is "upside down" and j-faces of P correspond to (d-1-j)-faces of P* (without proof)

30.11. (JK)

Remark about the graph of a 3-dimensional dual polytope
Simplicial polytope, simple polytope, examples
The f-vector of a polytope, estimates on the number of faces of 3-dimensional polytopes, definition of neighborly polytopes
Moment curve, cyclic polytope, Gale's evenness criterion characterizing facets of the cyclic polytope
Exact number of facets of the d-dimensional cyclic polytope with n vertices
The upper bound theorem (only statement)

7.12. (JK)

An asymptotic version of the upper bound theorem (proof only for simplicial polytopes)
Voronoi diagram, regions (cells), faces
Delaunay triangulation, properties without proofs
Voronoi diagrams in R^d are projections of polyhedra in R^{d+1} (proof using hyperplanes tangent to the unit paraboloid)
An asymptotic upper bound on the number of faces of all regions of the Voronoi diagram

14.12. (JK)

Power diagrams, their properties without proofs
Arrangement of hyperplanes in R^d, cells, faces
Simple arrangement of hyperplanes, the exact number of cells of a simple arrangement of n hyperplanes (proof by double induction on n, d)
Second proof for the exact number of cells of a simple arrangement of n hyperplanes (by induction on d)
Arrangement of arbitrary subsets of R^d
A zero set of a polynomial in d variables

21.12. (JK)

Arrangement of zero sets of polynomials, Milnor–Thom theorem, asymptotic and more precise version (no proof)
Sign patterns of a collection of polynomials
Arrangement of pseudolines, wiring diagram, allowable sequence, isomorphic arrangements
Stretchability of pseudoline arrangements, asymptotic numbers of pseudoline and line arrangements (using the Milnor–Thom theorem) imply that almost all pseudoline arrangements are not stretchable
Remark about the algorithmic complexity of stretchability

4.1. (PV)

Levels in an arrangement of n hyperplanes, Clarkson's theorem on levels (statement), a construction showing that the upper bound is asymptotically tight (outline)
Proof of Clarkson's theorem on levels for d=2 for simple arrangements
The zone theorem, easy proof of the bound of order O(n^d) (outline), easy proof of the lower bound of order \Omega(n^{d-1}) (outline)
proof of the zone theorem for d=2

Temata prednasek: