Mathematical Sciences
Sub-communities within this community
Collections in this community
Recent Submissions
-
Tasić, Jelena (Beograd , 2026)[more][less]
Abstract: This dissertation examines the p-next center problem (PNCP) and three of its variants that have not previously been addressed in the literature. All the problems consi- dered involve determining locations for establishing service centers, focusing on the user in the most unfavorable position. In practice, this may be the user who is farthest from their assigned health clinic, or the customer farthest from the local market. The objective is to ensure that these users travel the shortest possible distance to a service center. Since these are NP-hard problems, standard solvers such as CPLEX are unable to provide optimal, or often even feasible, solutions for larger instances. The p-next center problem reflects the realistic possibility that one or more centers may suddenly fail. In such cases, users assigned to a closed center are redirected to the (nearest) backup center, and the goal is to determine the locations for centers so as to minimize the maximum of all total distances traveled by users. In this dissertation, a skewed variable neighborhood search method (SVNS) is proposed for solving the p-next center problem, which incorporates a fast interchange heuristic within the local search phase. The method is tested on the well-known OR-LIB set of instances containing up to 900 nodes, and the results are compared with the best results from the literature. As an extension of the previous problem, the concept of facilitated communication be- tween centers is considered, and the p-next center problem with a discount factor (ωPNCP) is defined to incorporate this idea. A mathematical formulation of the problem is provided and solved using the CPLEX solver. A basic variable neighborhood search (BVNS) method is proposed for solving the problem, and the potential benefits achievable through enhanced communication between centers are analyzed on a set of instances from the literature that include up to 900 nodes. To further adapt the p-next center problem to practical needs, the conditional p-next center problem (CPNCP) is defined. This problem is applicable to the expansion of existing business networks while retaining existing centers where there is a possibility of sudden center failures. A mathematical model is proposed, instances with up to 900 nodes are generated, and the problem instances are solved using the CPLEX solver. A variable neighborhood search method is proposed for solving this problem. Two approaches to business network expansion are analyzed, along with potential long-term savings that can be achieved by their application. The maximal covering p-next center problem (MCPNCP) with binary and partial cove- rage is defined. The objective is to maximize the total demand of users that are covered, that is, users whose distance to their backup center does not exceed a given radius. Two mathematical models are proposed. Instance with up to 400 nodes are generated and the proposed models are compared using the CPLEX solver. A skewed variable neighborhood search method is proposed for solving the problem, and the results are compared with those obtained by the CPLEX solver. URI: http://hdl.handle.net/123456789/5803 Files in this item: 1
Jelena_Tasic_disertacija.pdf ( 6.076Mb ) -
Vidojević, Sonja (, 2012)[more][less]
Abstract: Interplanetary electron beams, produced by solar flares, are unstable in the solar wind and generate Langmuir waves at the local plasma frequency, fp. These waves are then converted into the so-called type III radio bursts which are freely propagating electromagnetic emissions at fp or its harmonic. The type IIIs are therefore observed as drifting emissions from high to low frequencies, in the kilometric wavelengths range. Since the first theoretical explanation by Ginzburg and Zhelezniakov (1958a), several refined models have attempted to describe in details the physical processes at the origin of type III bursts. The mechanisms of “Bump-on-tail” instabilities, Langmuir waves generations, conversion of these Langmuir waves into radio emissions throughout nonlinear wave-wave interactions etc, have been studied in detail. Of particular interest from the observational point of view are the so called in situ type III bursts for which the electron beam, at the origin of the emission and traveling along open interplanetary magnetic field lines, is observed directly in situ by a spacecraft, together with the local Langmuir waves and the resulting radio emissions. Until now only a few of these in situ type IIIs have been reported in the literature. The first research study performed in this thesis was to examine the first 16 years of radio, waves and particles data recorded by the Wind spacecraft in the Solar Wind and to look for in situ type IIIs. Applying rigorous and careful criteria, this examination has yielded to a data set of 36 high-quality events. With such a numerous data set, which is statistically representative of the studied phenomenon, it is now possible to constrain obser- vationally and with a better confidence the type III generation models. After having built our statistical dataset, we have studied, for each of the events, the precise shapes of the Langmuir wave power distributions, observed in the spectral domain. We have fitted these observed distributions by a Pearson’s system of probability distributions and have shown that the probability distributions of the logarithm of the Langmuir waves power spectral density belong to three “main” types of Pearson’s probability distributions: type I, type IV and type VI. In addition we have modeled the effects of the instrumental integration time of the Wind radio receivers on the observed Langmuir wave power distributions. By combining our observations with our models we have shown that it was not possible to con- clude definitively, that the distribution of the Langmuir waves energy in the real temporal domain is lognormal, as it is predicted in some theories as the Stochastic Growth Theory by Robinson (1992). In the last part of the thesis, we have shown how our high-quality data set of 36 in situ type III events can be used for further studies that could allow to constrain the theoretical models even better. For instance we have investigated the correlation between the Langmuir waves power and the energy of impulsive electron or with the power of the radio emissions themselves. URI: http://hdl.handle.net/123456789/5802 Files in this item: 1
Disertacija_Sonja.pdf ( 4.444Mb ) -
Šandrih, Branislava (Beograd , 2020)[more][less]
Abstract: The main goal of this dissertation is to put different text classification tasks in the same frame, by mapping the input data into the common vector space of linguistic attributes. Subsequently, several classification problems of great importance for natural language processing are solved by applying the appropriate classification algorithms. The dissertation deals with the problem of validation of bilingual translation pairs, so that the final goal is to construct a classifier which provides a substitute for human evalu- ation and which decides whether the pair is a proper translation between the appropriate languages by means of applying a variety of linguistic information and methods. In dictionaries it is useful to have a sentence that demonstrates use for a particular dictio- nary entry. This task is called the classification of good dictionary examples. In this thesis, a method is developed which automatically estimates whether an example is good or bad for a specific dictionary entry. Two cases of short message classification are also discussed in this dissertation. In the first case, classes are the authors of the messages, and the task is to assign each message to its author from that fixed set. This task is called authorship identification. The other observed classification of short messages is called opinion mining, or sentiment analysis. Starting from the assumption that a short message carries a positive or negative attitude about a thing, or is purely informative, classes can be: positive, negative and neutral. These tasks are of great importance in the field of natural language processing and the proposed solutions are language-independent, based on machine learning methods: sup- port vector machines, decision trees and gradient boosting. For all of these tasks, a demonstration of the effectiveness of the proposed methods is shown on for the Serbian language. URI: http://hdl.handle.net/123456789/5801 Files in this item: 1
Disertacija.pdf ( 9.053Mb ) -
Perić, Milan (Beograd , 2021)[more][less]
Abstract: This thesis presents a method for calculating the polynomial entropy of the topolog- ical dynamic system with finitely many non-wandering points. A special coding is adapted for such systems. Thanks to this coding the polynomial entropy can be bounded by the number of specific mutually singular points in the closures of stable manifolds of non-wandering points. This method was applied to Morse gradient systems. It is shown that the polynomial entropy of the Morse gradient system is bounded by n(F ) − 1, where n(F ) is the number of different Morse indices of critical points of the Morse function F. If Morse gradient systems on mani- folds of dimension n has only critical points of indices 0 and n, it is proved that the polynomial entropy is equal to 1, and if the system has critical points of indices 0, n/2 and n, it is proved that polynomial entropy is equal to 2. The polynomial entropy for different parameter values in logistic map has also been calculated, and it has been shown that the polynomial entropy distinguishes the systems of low complexity with drastically different behaviours, which cannot be distinguished by the topological entropy. The example of the homeomorphism of the con- nected compact metric space that is not equicontinuous and with vanishing polynomial entropy is also given. URI: http://hdl.handle.net/123456789/5800 Files in this item: 1
Disertacija_12349.pdf ( 731.8Kb ) -
Jovanović Spasojević, Tanja (Beograd , 2022)[more][less]
Abstract: In this thesis, subjects of consideration are the embeddings theorems of weighted Bergman spaces in Lp-spaces, as well as embeddings theorems of harmonic mixed norm spaces. The first part of the thesis generalizes the theorems of embeddings Bergman spaces into Lp(μ)-spaces, where μ is a Borel measure on a given domain. They have been earlier studied on domains such as unit ball and upper half-space. Generalization refers to bounded domains Ω ⊂ Rn with C1 boundary. This embedding will be valid to any p > 0, whenever the measure of the spaces Lp satisfies the Carledon condition. Reverse the direction will be valid only in case if p > 1 + α+2 n−2 . The second part of the dissertation also generalizes the embeddings theorems of mixed norm spaces of harmonic functions on a unit ball, where the generalization is applied to the domain Ω ⊂ Rn with C1 boundary. However, in addition we are obtaining another important result relating to the limitation of the maximum operators in the mixed norm on the general domain for the class of QNS functions. URI: http://hdl.handle.net/123456789/5799 Files in this item: 1
Disertacija_13689.pdf ( 1.643Mb )