Refine
Document Type
- Article (119) (remove)
Institute
Keywords
- 1996 (51)
- Bauhaus-Kolloquium (51)
- Weimar (51)
- Architektur (32)
- Technologie (21)
- Fiktion (13)
- Modellierung (10)
- CAD (6)
- Bauwerk (5)
- Finite-Elemente-Methode (4)
Year of publication
- 1997 (119) (remove)
Gegenstand der Betrachtung ist ein spezielles Tourenproblem der Schüttgutoptimierung. Man stelle sich als Realitätsbezug ein Transportunternehmen vor, das eine Anzahl von gleichartigen Fahrzeugen in einem Fuhrpark stationiert hat. Vorgegebene Mengen von Schüttgut müssen von einer Kiesgrube zu mehreren Baustellen transportiert werden. Dabei sind Be- und Entladezeiten, unterschiedliche mittlere Geschwindigkeiten für Leer- bzw. Lastfahrten und Schichtzeiten zu berücksichtigen. Gesucht ist eine optimale Anzahl von einzusetzenden Fahrzeugen und die zugehörigen Tourenpläne mit dem Ziel der Minimierung der Transportkosten unter Beachtung der Lieferverträge. Die Lösung des Problems erfolgt in zwei Phasen. Zuerst wird die Frage geklärt, wieviele Fahrzeuge bei minimalem Kostenniveau einzusetzen und welche zu den Leerfahrten gehörigen Teilstrecken wie oft zu befahren sind. Anschließend wird mit heuristischen Verfahren versucht, die Menge der zu fahrenden Teilstrecken so auf die Fahrzeuge aufzuteilen, daß für jedes Fahrzeug eine zulässige Tour entsteht. Zur komplexen Lösung einer denkbaren Aufgabe liegt ein Programm in der Programmiersprache PASCAL vor. Die erzielten numerischen Resultate belegen, daß auch für Probleme größerer Dimension eine Optimallösung oder sehr gute Näherungen in vernünftiger Zeit gefunden werden.
Priority-rule methods for approximately minimizing the duration of a project subject to minimal and maximal time lags between the activities of the project and limited availability of renewable resources are considered. Such a project can be modelled by a cyclic activity-on-node network. Two generation schemes for constructing feasible schedules are discussed: the serial and parallel schemes. Two different kinds of heuristic procedures are proposed. The sequential or direct method processes the activities or respectively nodes of the project network one after another without considering the strong components separately. The contraction method uses a bottom-up technique. First, a feasible subschedule is determined for each strong component. Second, each strong component is replaced by a single node and the resulting acyclic network is treated by the direct method. In conclusion, some results from an experimental performance analysis of the heuristics are given using a new network generator.
We consider the standardization problem (SP) which can be formulated as follows. It is known demand bi in each type i in {1, 2, ..., n} of items. Production of yi items of the ith type brings a profit fi (yi), where fi is a nondecreasing concave function for each i in {1, 2, ..., n}.It is necessary to satisfy the demand and to maximize the total profit provided that there exist >standardization possibilities< . These possibilities means that some types of items can be replaced by some another types. We introduce generalized standardization problem (GSP) in which titems demand is given as the set of admissible demand vectors. We show that GSP and SP are special cases of the resource allocation problem over a network polymatroid. Ibasing on this observation we propose a polynomial time solution algorithm for GSP and SP.
The idea of representing urban structure and various communication systems (water and energy supply, telephone and cable TV networks) as fractal objects is not absolutely new. However, known works, devoted to this problem use models and approaches from fractal physics. For example, to simulate urban growth Diffusion Limited Aggregation (DLA) model and Dielectric Breakdown (DB) model are used. This study introduces a different approach. Net structure of communication system is described by a graph of special type called regular G(l,r,n)-graph. Authors provide description of such graph, develop iterative process for its generation and prove its self-similarity, i.e. that every regular graph is a pre-fractal. After the infinite number of steps this process generates a fractal. The devised algorithm for generation and grathical representation of regular G(l,r,n)-graphs with different values of l,r and n has been programmed to receive computer simulations. For optimal graphic presentation of pre-fractals the Optimal Space Ordering method was suggested. It is based on the minimization of the >graph energy< value about vertices' coordinates. The effective procedure for optimization was developed that takes into account specific properties of graph energy as objective function For the fractal graph introduced the Hausdorff-Besikovich and similarity dimensions were calculated. It has been shown that >graph energy< is directly related to the graph's fractal properties. For G(3,3,n) and G(4,4,n) graphs fractal dimensions calculated by different methods are the same (D=1,5 and D=2 respectively), while topological dimension of both graphs is 1.
A multicriterial statement of the above mentioned problem is presented. It differes from the classical statement of Spanning Tree problem. The quality of solution is estimated by vector objective function which contains weight criteria as well as topological criteria (degree and diameter of tree). Many real processes are not determined yet. And that is why the investigation of the stability is very important. Many errors are connected with calculations. The stability analysis of vector combinatorial problems allows to discover the value of changes in the initial data for which the optimal solution is not changed. Furthermore, the investigation of the stability allows to construct the class of the problems on base of the one problem by means of the parameter variations. Analysis of the problems with belong to this class allows to obtaine axact and adecuate discription of model
This work was partially supported by DAAD, Fundamental Researches Foundation of Belarus and International Soros Science Education Program We consider a vector discrete optimization problem on a system of non- empty subsets (trajectories) of a finite set. The vector criterion of the pro- blem consists partial criterias of the kinds MINSUM, MINMAX and MIN- MIN. The stability of eficient (Pareto optimal, Slater optimal and Smale op- timal) trajectories to perturbations of vector criterion parameters has been investigated. Suficient and necessary conditions of eficient trajectories local stability have been obtained. Lower evaluations of eficient trajectories sta- bility radii, and formulas in several cases, have been found for the case when l(inf) -norm is defined in the space of vector criterion parameters.
The theory of random matrices, or random matrix theory, RMT in what follows, has been developed at the beginning of the fties to describe the sta- tistical properties of energy levels of complex quantum systems, [1], [2], [3]. In the early eighties it has enjoyed renewed interest since it has been recognized as a very useful tool in the study of numerous physical systems. Specically, it is very useful in the analysis of chaotic quantum systems. In fact, in the last years many papers appeared about the problem of quantum chaos which implies the quantization of systems whose underlying classical dynamics is irregular (i.e. chaotic). The simplest models considered in this eld are billi- ards of various shapes. From the the classical point of view, a point particle in a 2-dimensional billiard displays regular or irregular motion depending on the shape of the billiard; for instance motion in a rectangular or circular billi- ard is regular thanks to the symmetries of the boundary. On the other hand, billiards of arbitrary shapes imply chaotic motion, i.e. exponential diver- gence of initially nearby trajectories. In order to study quantum billiards we have to consider the Schroedinger equation in various 2-dimensional domains. The eigenvalues of the Schroedinger equation represent the allowed energy levels of our quantum particle in the billiard under consideration, while the eigenfunction norms represent the probability density of nding the particle in a certain position. The question of quantum chaos is whether the charac- ter of the classical motion (regular or chaotic) can in uence some properties
With the help of functional analytical methods complex analysis is a powerful tool in treating non-linear first-order partial differential equations in the plane. Some of the most important of these equations are the Beltrami equations. This is due to the fact that the theory of Beltrami systems is related to many problems of geometry and analysis, like non-linear subsonic two-dimensional hydrodynamics, problems of conformal and quasiconformal mappings of two-dimensional Riemannian manifolds, or complex analytic dynamics. The theory of Beltrami equations is strongly connected with the -operator. This singular integral operator is immediately recognized as two-dimensional Hilbert-transform, known also under the name of integral operator with Beurling kernel, acting as an isometry of L2(C) onto L2(C). In hypercomplex function theory the Beltrami equations have not yet this importance, but nevertheless, they are a basic condition for the transfer of complex methods and efforts for solving partial differential equations, especially of non-linear type, to the spatial case. Here we deal with hypercomplex Beltrami systems. For this we restrict ourselves to the quaternionic case, but without any loss of generality. We will show how a spatial generalization of the complex -operator can be used to solve systems of non-linear partial differential equations, in particular different types of spatial Beltrami systems. Also, the for practical purposes so important norm estimates will be derived. Some of our results are stronger as known results in the complex case, but they are applicable in the complex situation, too.
The aim of this talk is to show that the methods used by Métivier and Lapidus to study the eigenvalue distribution of elliptic operators (e.g., of the Dirichlet Laplacian) can be adapted to the study of the similar problem for the Stokes operator. In this way we get asymptotic formulae for the eigenvalues of the latter operator even in the case when the underlying domain has an extremely irregular (fractal) boundary. In the case the boundary is not that irregular (e.g., when it is Lipschitz) the estimates we obtain are much better than the ones we can find in the current literature.
Bauwerke sind in ihrer Betriebszeit vielen nutzungseinschränkenden Einflüssen ausgesetzt. Die dadurch erforderliche Instandhaltung dient zur Gewährleistung und zur Erhöhung der geplanten Nutzungsfähigkeit sowie Dauerhaftigkeit von Bauwerken. Sie spielt eine immer größere Rolle im Bauwesen. Die Kosten der Bauwerksinstandhaltung betragen je nach Art des Bauwerkes in Deutschland pro Jahr ca. 1-6% des Wiederbeschaffungswertes. Die Reduzierung des Instandhaltungsaufwandes durch Technik- und Management-Maßnahmen ist daher wirtschaftlich sinnvoll. Der Einsatz moderner Informations- und Kommunikationstechnologie auf dem Gebiet der Bestandsaufnahme und -analyse ist erforderlich, um einerseits die Bearbeitung von multimedialen Informationen über Ist- und Soll-Zustände von Bauwerken effektiver durchzuführen und um andererseits die Analyse von Schäden im Sinne einer Entscheidungshilfe zu unterstützen. Durch den Einsatz der WWW-Technologie kann die Bearbeitung auch verteilt im Datennetz über ferne Rechner hinweg erfolgen. Im Beitrag werden die Konzeptionierung und Implementierung des WWW-fähigen DV-Systems BINAS zur Unterstützung der Bestandsaufnahme und -analyse sowie die dafür erforderlichen Methoden und Werkzeuge vorgestellt.