### Refine

#### Year of publication

- 2001 (18) (remove)

#### Document Type

- Preprint (18) (remove)

#### Language

- English (18) (remove)

#### Keywords

- heat equation (2)
- stationary radiative transfer equation (2)
- Associative Memory Problem (1)
- Capacity (1)
- Earth' (1)
- Feed-forward Networks (1)
- Hyperbolic Conservation (1)
- Industrial Applications (1)
- Learnability (1)
- Navier Stokes equation (1)

#### Faculty / Organisational entity

- Fachbereich Mathematik (18) (remove)

The anchored hyperplane location problem is to locate a hyperplane passing through some given points P IR^n and minimizing either the sum of weighted distances (median problem), or the maximum weighted distance (center problem) to some other points Q IR^n . If the distances are measured by a norm, it will be shown that in the median case there exists an optimal hyperplane that passes through at least n - k affinely independent points of Q, if k is the maximum number of affinely independent points of P. In the center case, there exists an optimal hyperplane which isatmaximum distance to at least n - k + 1 affinely independent points of Q. Furthermore, if the norm is a smooth norm, all optimal hyperplanes satisfy these criteria. These new results generalize known results about unrestricted hyperplane location problems.

A natural extension of point facility location problems are those problems in which facilities are extensive, i.e. those that can not be represented by isolated points but as some dimensional structures such as straight lines, segments of lines, polygonal curves or circles. In this paper a review of the existing work on the location of extensive facilities in continuous spaces is given. Gaps in the knowledge are identified and suggestions for further research are made.

Abstract: Evacuation problems can be modeled as flow problems in dynamic networks. A dynamic network is defined by a directed graph G = (N,A) with sources, sinks and non-negative integral travel times and capacities for every arc (i,j) e A. The earliest arrival flow problem is to send a maximum amount of dynamic flow reaching the sink not only for the given time horizon T, but also for any time T' < T . This problem mimics the evacuation problem of public buildings where occupancies may not known. For the buildings where the number of occupancies is known and concentrated only in one source, the quickest flow model is used to find the minimum egress time. We propose in this paper a solution procedure for evacuation problems with a single source of the building where the occupancy number is either known or unknown. The possibility that the flow capacity may change due to the increasing of smoke density or fire obstructions can be mirrored in our model. The solution procedure looks iteratively for the shortest conditional augmenting path (SCAP) from source to sink and compute the time intervals in which flow reaches the sink via this path.

Given a railway network together with information on the population and their use of the railway infrastructure, we are considering the e ffects of introducing new train stops in the existing railway network. One e ffect concerns the accessibility of the railway infrastructure to the population, measured in how far people live from their nearest train stop. The second effect we study is the change in travel time for the railway customers that is induced by new train stops. Based on these two models, we introduce two combinatorial optimization problems and give NP-hardness results for them. We suggest an algorithmic approach for the model based on travel time and give first experimental results.

In this article, we investigate the maximum entropy moment closure in gas dynamics. We show that the usual choice of polynomial weight functions may lead to hyperbolic systems with an unpleasant state space: equilibrium states are boundary points with possibly singular fluxes. In order to avoid singularities, the necessary arises to find weight functions which growing sub-quadratically at infinity. Unfortunately, this requirement leads to a conflict with Galilean invariance of the moment systems because we can show that rotational and translational invariant, finite dimensional function spaces necessarily consist of polynomials.

Abstract: The basic concepts of selective multiscale reconstruction of functions on the sphere from error-affected data is outlined for scalar functions. The selective reconstruction mechanism is based on the premise that multiscale approximation can be well-represented in terms of only a relatively small number of expansion coefficients at various resolution levels. A new pyramid scheme is presented to efficiently remove the noise at different scales using a priori statistical information.

The purpose of satellite-to-satellite tracking (SST) and/or satellite gravity gradiometry (SGG) is to determine the gravitational field on and outside the Earth's surface from given gradients of the gravitational potential and/or the gravitational field at satellite altitude. In this paper both satellite techniques are analysed and characterized from mathematical point of view. Uniqueness results are formulated. The justification is given for approximating the external gravitational field by finite linear combination of certain gradient fields (for example, gradient fields of single-poles or multi-poles) consistent to a given set of SGG and/or SST data. A strategy of modelling the gravitational field from satellite data within a multiscale concept is described; illustrations based on the EGM96 model are given.

Integral equations on the half of line are commonly approximated by the finite-section approximation, in which the infinite upper limit is replaced by apositie number called finite-section parameter. In this paper we consider the finite-section approximation for first kind intgral equations which are typically ill-posed and call for regularization. For some classes of such equations corresponding to inverse problems from optics and astronomy we indicate the finite-section parameters that allows to apply standard regularization techniques. Two discretization schemes for the finite-section equations ar also proposed and their efficiency is studied.

By means of the limit and jump relations of classical potential theory the framework of a wavelet approach on a regular surface is established. The properties of a multiresolution analysis are verified, and a tree algorithm for fast computation is developed based on numerical integration. As applications of the wavelet approach some numerical examples are presented, including the zoom-in property as well as the detection of high frequency perturbations. At the end we discuss a fast multiscale representation of the solution of (exterior) Dirichlet's or Neumann's boundary-value problem corresponding to regular surfaces.