02374ntm a22003017a 4500003000800000005001700008008004100025040000800066100002100074245004700095260002200142500001100164505001300175505002000188505002500208505001900233505001400252505003800266505001600304505007900320505004900399505003200448505002400480505001700504505001000521520149700531856004402028AT-ISTA20190813103052.0180627s2017 au ||||| m||| 00| 0 eng d cIST aNikitenko, Anton aDiscrete Morse theory for random complexes bIST Austriac2017 aThesis aAbstract aAcknowledgments aList of publications a1 Introduction a2 Results a3 Blaschke- Petkantschin formulas a4 Constants a5 Poisson-Delaunay, Poisson-Čech and weighted Poisson-Delaunay complexes a6 Poisson-Delaunay complexes of higher order a7 Random inscribed polytops a8 Future directions aBibliography aIndex aThe main objects considered in the present work are simplicial and CW-complexes with vertices forming a random point cloud.
In particular, we consider a Poisson point process in R^n
and study Delaunay and Voronoi complexes of the first and higher orders and weighted Delaunay complexes obtained as sections of Delaunay complexes, as well as the Čech complex.
Further, we examine theDelaunay complex of a Poisson
point process on the sphere S^n, as well as of a uniform point cloud, which is equivalent to the convex hull, providing a connection to the theory of random polytopes.
Each of the complexes in question can be endowed with a radius function,
which maps its cells to the radii of appropriately chosen circumspheres,
called the radius of the cell.
Applying and developing discrete Morse theory for these functions, joining it together with probabilistic and sometimes analytic machinery, and developing several integral geometric tools, we aim at getting the distributions of circumradii of typical cells.
For all considered complexes, we are able to generalize and obtain up to constants
the distribution of radii of typical intervals of all types.
In low dimensions the constants
can be computed explicitly, thus providing the explicit
expressions for the expected numbers of cells.
In particular, it allows to find the expected density of simplices
of every dimension for a Poisson point process in R^4,
whereas the result for R^3 was known already in 1970's. uhttps://doi.org/10.15479/AT:ISTA:th_873