Browsing Doctoral Dissertations by Title

Pavlović, Marko (, 2017)[more][less]
Abstract: Supernova (SN) explosions disperse the different heavy elements across the Uni verse. These elements are the building blocks which make up the world around and inside us. Supernova remnants (SNRs) are extraordinary astronomical objects that are also of high scientific interest, because they provide insights into aforementioned supernova explosion mechanisms, and because they are important sources of Galac tic cosmic rays (CRs). Radio observations are among the oldest means to study these objects. The radio luminosity and spectra of SNRs, especially young ones, requires active acceleration of electrons by the SNR shocks. In this doctoral dissertation, radio evolution of SNRs is investigated by using threedimensional hydrodynamic modelling and nonlinear diffusive shock acceleration of CRs in SNRs. Hydrodynamic simulations, developed and adopted in this dissertation, allow us to explicitly account for the shock modification by CRs. We also include consistent numerical treatment of magnetic field amplification (MFA) due to CR resonant and nonresonant streaming instabilities. We modelled the peculiar nature of radio evo lution of the youngest known Galactic SNR G1.9+0.3 and concluded that increasing radio emission is a common occurrence among very young SNRs. Our model ena bled us to make important conclusions about the present and predictions about the future properties of radio emission from this SNR. We also developed more general model of the radio evolution of SNRs, by performing simulations for wide range of the relevant physical parameters, such as the ambient density, the supernova ex plosion energy, the acceleration efficiency and the MFA efficiency. We confirm the reliability of our radio evolutionary tracks on a observation sample consisting of Galactic and extragalactic SNRs. This dissertation also deals with one of the most important questions surroun ding our current understanding of the magnetic fields in SNRs. We conclude that equipartition is a justified assumption especially between the CR electrons and the magnetic fields in evolved SNRs, in the SedovTaylor phase of evolution. Our work also offers a possible explanation how can equipartition between CRs and magnetic field in the interstellar medium be achieved. Type of modeling, presented in this thesis, is expected to be a useful tool for fu ture observers working on powerful radio telescopes such as ALMA, MWA, ASKAP, SKA and FAST. Simulations should provide important information about the evolu tionary stage of the observed SNRs, as well as to characterize the physical conditions in the shocks where the relativistic particles are accelerated. Simulations could help us to predict the science output of future large scale surveys, as well as to explain new, often unexpected results obtained by observations. URI: http://hdl.handle.net/123456789/4759 Files in this item: 1
MPavlovic.pdf ( 14.32Mb ) 
Pavlović, Marko (Beograd , 2018)[more][less]
Abstract: Supernova (SN) explosions disperse the diﬀerent heavy elements across the Uni verse. These elements are the building blocks which make up the world around and inside us. Supernova remnants (SNRs) are extraordinary astronomical objects that are also of high scientiﬁc interest, because they provide insights into aforementioned supernova explosion mechanisms, and because they are important sources of Galac tic cosmic rays (CRs). Radio observations are among the oldest means to study these objects. The radio luminosity and spectra of SNRs, especially young ones, requires active acceleration of electrons by the SNR shocks. In this doctoral dissertation, radio evolution of SNRs is investigated by using threedimensional hydrodynamic modelling and nonlinear diﬀusive shock acceleration of CRs in SNRs. Hydrodynamic simulations, developed and adopted in this dissertation, allow us to explicitly account for the shock modiﬁcation by CRs. We also include consistent numerical treatment of magnetic ﬁeld ampliﬁcation (MFA) due to CR resonant and nonresonant streaming instabilities. We modelled the peculiar nature of radio evo lution of the youngest known Galactic SNR G1.9+0.3 and concluded that increasing radio emission is a common occurrence among very young SNRs. Our model ena bled us to make important conclusions about the present and predictions about the future properties of radio emission from this SNR. We also developed more general model of the radio evolution of SNRs, by performing simulations for wide range of the relevant physical parameters, such as the ambient density, the supernova ex plosion energy, the acceleration eﬃciency and the MFA eﬃciency. We conﬁrm the reliability of our radio evolutionary tracks on a observation sample consisting of Galactic and extragalactic SNRs. This dissertation also deals with one of the most important questions surroun ding our current understanding of the magnetic ﬁelds in SNRs. We conclude that equipartition is a justiﬁed assumption especially between the CR electrons and the magnetic ﬁelds in evolved SNRs, in the SedovTaylor phase of evolution. Our work also oﬀers a possible explanation how can equipartition between CRs and magnetic ﬁeld in the interstellar medium be achieved. Type of modeling, presented in this thesis, is expected to be a useful tool for fu ture observers working on powerful radio telescopes such as ALMA, MWA, ASKAP, SKA and FAST. Simulations should provide important information about the evolu tionary stage of the observed SNRs, as well as to characterize the physical conditions in the shocks where the relativistic particles are accelerated. Simulations could help us to predict the science output of future large scale surveys, as well as to explain new, often unexpected results obtained by observations. URI: http://hdl.handle.net/123456789/4736 Files in this item: 1
MPavlovic.pdf ( 14.32Mb ) 
Spasić, Mirko (Beograd , 2020)[more][less]
Abstract: The query containment problem is a very important computer science problem,originally defined for relational queries. With the growing popularity of theSPARQLquerylanguage, it became relevant and important in this new context, too. This thesis introducesa new approach for solving this problem, based on a reduction to satisfiability in first orderlogic. The approach covers containment underRDF SCHEMAentailment regime, and it candeal with the subsumption relation, as a weaker form of containment. The thesis provessoundness and completeness of the approach for a wide range of language constructs. It alsodescribes an implementation of the approach as an open source solverSPECS. The experimental evaluation on relevant benchmarks shows thatSPECSis efficient and comparing tostateoftheart solvers, it gives more precise results in a shorter amount of time, while supporting a larger fragment ofSPARQLconstructs. An application of query language modeling canbe useful also along refactoring of database driven applications, where simultaneous changesthat include both a query and a host language code are very common. These changes canpreserve the overall equivalence, without preserving equivalence of these two parts consideredseparately. Because of the ability to guarantee the absence of differences in behavior betweentwo versions of the code, tools that automatically verify code equivalence have great benefitsfor reliable software development. With this motivation, a custom firstorder logic modelingof SQL queries is developed and described in the thesis. It enables an automated approachfor reasoning about equivalence ofC/C++programs with embedded SQL. The approach is implemented within a publicly available and open source framework SQLAV. URI: http://hdl.handle.net/123456789/5095 Files in this item: 1
doktoratMirkoSpasic.pdf ( 1.258Mb ) 
Marković, Zoran (Pennsylvania)[more][less]
Abstract: The results from this thesis are obtained by using notions and procedures which are wellknown in Kripke structures in the first place, together with some other constructions. They might provides insights about intuitionistic formal theories analogous to insights about classical logic provided by results of classical model theory. The thesis consists of three chapters. The definitions concerning syntax of the first order intuitionistic logic, the definitions and theorems about Kripke structures, Hayting algebras and saturated theories are given in Chapter 1. In the first part of the next chapter a few results about the connection between forcing and classical satisfaction relation are proved. In the second part of that chapter three alternatives of the antecedent of the omitting type theorem are presented, and an omitting types theorem is proved. It is important that there are many applications of that theorem. In Chapter 3 the following two kinds of products are considered: prime products of saturated theories and ultra products and reduced products of Kripke structures. In the first part of that chapter the following property is proved: a simple analogue of ultraproduct construction can be defined in terms of saturated theories. The important result from the second part of Chapter 3 is that the class of formulas preserved under reduced products is much broader than the class of formulas which are intuitionistically equivalent to Horn formulas. URI: http://hdl.handle.net/123456789/317 Files in this item: 1
phdZoranMarkovic.pdf ( 9.114Mb ) 
Dražić, Zorica (Beograd , 2014)[more][less]
Abstract: The Variable neighborhood search method proved to be very successful for solving discrete and continuous optimization problems. The basic idea is a systematic change of neighborhood structures in search for the better solution. For optimiza tion of multiple variable functions, methods for obtaining the local minimum starting from certain initial point are used. In case when the continuous function has many local minima, nding the global minimum is usually not an easy task since the obta ined local minima in most cases are not optimal. In typical implementations with bounded neighborhoods of various diameters it is not possible, from arbitrary point, to reach all points in solution space. Consequently, the strategy of using the nite number of neighborhoods is suitable for problems with solutions belonging to some known bounded subset of IRn. In order to overcome the previously mentioned limitation the new variant of the method is proposed, Gaussian Variable neighborhood search method. Instead of de ning the sequence of di erent neighborhoods from which the random point will be chosen, all neighborhoods coincide with the whole solution space, but with di e rent probability distributions of Gaussian type. With this approach, from arbitrary point another more distant point is theoretically reachable, although with smaller probability. In basic version of Variable neighborhood search method one must de ne in advance the neighborhood structure system, their number and size, as well as the type of random distribution to be used for obtaining the random point from it. Gaussian Variable neighborhood search method has less parameters since all the neighborhoods are theoretically the same (equal to the solution space), and uses only one distribution family  Gaussian multivariate distribution with variable dispersion. File transfer scheduling problem (FTSP) is an optimization problem widely appli cable to many areas such as Wide Area computer Networks (WAN), Local Area Ne v tworks (LAN), telecommunications, multiprocessor scheduling in a MIMD machines, task assignments in companies, etc. As it belongs to the NPhard class of problems, heuristic methods are usually used for solving this kind of problems. The problem is to minimize the overall time needed to transfer all les to their destinations for a given collection of various sized les in a computer network, i.e. to nd the le transfer schedule with minimal length. In order to obtain the exact solution of the FTS problem, integer linear pro gramming formulations are proposed and their correctness is proved. In this way optimal solutions can be found for small and medium size test instances. For large test instances, the Variable neighborhood search method is proposed using the "permutation" representation and typical neighborhood structures. More over, the same method is used for obtaining the upper bounds of the solutions which are used in proposed integer linear programming formulations. For obtaining be tter solutions in the small neighborhood of the current solution, three di erent local search procedures are implemented: 2swap, 2swap adjacent and variable neighbo rhood descent. In order to apply the continuous optimization methods for solving FTSP, the weighted solution representation is developed. Such representation enables the co ntinuous optimization methods to be used, which do not require the di erentiability of objective function. Since Gauss Variable neighborhood search method proved to be successful in continuous optimization problems, it was applied to FTSP. Pre viously described local search procedures can also be used with weighted solution representation. Using the proposed methods optimal solutions for all small and medium size test instances are found. For large size instances, which are beyond the reach of exact methods, metaheuristic methods obtained good solutions in reasonable time. URI: http://hdl.handle.net/123456789/4246 Files in this item: 1
phdDrazic_Zorica.pdf ( 4.739Mb ) 
Mihajlović, Borivoje (Belgrade , 1964)[more][less]
URI: http://hdl.handle.net/123456789/227 Files in this item: 1
phdBorivojeMihajlovic.pdf ( 2.509Mb ) 
Manojlović, Vesna (Beograd , 2008)[more][less]

Nikolić, Silvana (Belgrade)[more][less]

Marjanović, Miroslav (Belgrade)[more][less]

Kamberi, Qerim (Priština)[more][less]

Tanasijević, Ivana (Beograd , 2020)[more][less]
Abstract: The motivation for writing this doctoral dissertation is a multimedia collection that is the result of many years of field research conducted by researchers from the Institute for Balkan studies of the Serbian Academy of Sciences and Arts. The collection consists of materials in the form of recorded interviews, various recorded customs, associated textual descriptions (protocols) and numerous other documents.The subject of research of this dissertation is the study of possibilities and the development of new methods that could be used as a starting point in solving the problem of managing the intangible cultural heritage of the Balkans. The subtasksthat emerge in this endeavor are the development of adequate design and implementation of a multimedia database of intangible cultural heritage that would meet the needs of different types of users, automatic semantic annotation of protocols using natural language processing methods, as a basis for semiautomatic annotation of the multimedia collection, and successful search by metadata which comply with the CIDOC CRM standard, study of additional search possibilities of this collection in order to gain new knowledge, as well as development of selected methods.The main problem with the available methods is that there is still not enough developed infrastructure in the context of natural language processing, organization and management in the field of cultural heritage in the Balkans and especially for the Serbian language, which could be effectively used to solve the proposed problem.There is thus a strong need to develop methods to reach an appropriate solution.For the semiautomatic annotation of multimedia materials, automatic semantic annotation of the protocols associated with the materials was used. It was carriedout by methods of information extraction, recognition of named entities and topicextraction, using rulebased techniques with the help of additional resources suchas electronic dictionaries, thesauri and vocabularies from a specific domain.To classify textual protocols in relation to the topic, research was conducted onmethods that can be used to solve the problem of classifying texts in the Serbianlanguage, and a method was offered that is adapted to the specific domain beingprocessed (intangible cultural heritage), to the specific problems being solved (classification of protocols in relation to the topic) and to the Serbian language, as one of the morphologically rich languages.To work with spatial data, a spatial model has been developed that is suitable for displaying results on a map, as well as for creating spatial queries through an interactive graphical display of a map of locations.The results of experiments conducted on the developed methods show that the use of a rulebased approach in combination with additional language resources an dwith putting in a reasonable amount of effort gives very good results for the task of information extraction. An F measure of 0.87 was reached for the extraction of named entities, while an F measure of 0.90 was reached for the extraction of topics,which is in the range of measures from published research from similar problem sand domains.The results of the text classification indicate that the selected statistical methods of machine learning in their basic form when applied to the protocols, although generally successful, give a bad F measure, 0.44, while significant improvement is achieved with the use of semantic techniques, in which case an F measure of 0.88 is reached.Some of the results presented in this dissertation are contained in the papers[266], [265], [94], [264], [267], which have been published or accepted for publication.The conclusion drawn from the research is that to solve the given problem it is necessary to engage experts from several fields, that the needs of different groups of users are complex, which complicates the task of organizing and managing them ultimedia collection, that the domain of cultural heritage is very rich in semantics,that context plays a major role in the tasks of information extraction and text classification, and finally that for these tasks the developed rulebased methods of natural language processing as well as statistical techniques of machine learning prove to be successful. URI: http://hdl.handle.net/123456789/5093 Files in this item: 1
IvanaTanasijevicdr.pdf ( 4.674Mb ) 
Janković, Slobodan (Beograd , 1979)[more][less]

Marković, Marijan (Belgrade , 2013)[more][less]
Abstract: This work consists of three chapters. The first one contains some well known facts about Hardy classes of harmonic, analytic, and logarithmically subharmonic functions in the unit disk, as well as their applications. Then we briefly talk about the harmonic and minimal surfaces, the classical isoperimetric inequality, and the more recent results related to this inequality. One of the most elegant way to establish the isoperimetric inequality is via Carleman’s inequality for analytic functions in disks. In the second chapter we present the results from our recent work [29] for harmonic mappings of a disc onto a Jordan surface. In this chapter we establish the versions of classical theorems of Carath´eodory and Smirnov for mappings of the previous type. At the end of the head we apply these results to prove the isoperimetric inequality for Jordan harmonic surfaces bounded by rectifiable curves. In the third chapter, according to the author paper [35], we prove an inequality of the isoperimetric type, similar to Carleman’s, for functions of several variables. The first version of this inequality is for analytic functions in a Reinhardt domain. The second one concerns the functions that belong to Hardy spaces in polydiscs. URI: http://hdl.handle.net/123456789/2586 Files in this item: 1
Markovic_Marijan.pdf ( 709.3Kb ) 
Lazarević, Milan (Beograd , 2020)[more][less]
Abstract: ist of already known and recently established CauchySchwarz inequalities fore lementary operators,σelementary transformers and inner product type transformers iscomplemented by the next variant of CauchySchwarz inequality in Schattenvon Neumann ideals: ZΩAtA∗tdμ(t) 12q−12ZΩAtXBtdμ(t) ZΩB∗tBtdμ(t) 12r−12p6 ZΩA∗tAtdμ(t) 12qX ZΩBtB∗tdμ(t) 12rpfor allX2Cp(H)and for allp, q, r>1which satisfies2p=1q+1r,if families of operatorsfAtgt∈Ω,fA∗tgt∈Ω,fBtgt∈Ω,fB∗tgt∈Ωare strongly square integrabile, such thatRΩAtA∗tdμ(t)andRΩB∗tBtdμ(t)are (boundedly) invertible operators.Enabled by some additional conditions of commutativity and normality for operator familiesfAng∞n=1,fBng∞n=1,fAtgt∈ΩandfBtgt∈Ω,as well as by the degree ofpmodifications ofunitary invariant norms, some others variants of those inequalities will also be considered,including their applications to the certain problems in operator theory, norm inequalities forgeneralized function derivations of Pick operator values functions and operator values Fouriertransformations of complex measures, as well as some GrüssLandau type inequalities. URI: http://hdl.handle.net/123456789/5092 Files in this item: 1
disertacija_Lazarevic_Milan_dorada.pdf ( 1.868Mb ) 
Kovačević, Ilija (Belgrade)[more][less]

Gilezan, Koriolan (Belgrade)[more][less]

Jovanović, Boško (Belgrade)[more][less]

Hoxha, Isak (Priština)[more][less]

Mihajlović, Bojana (Beograd , 2016)[more][less]
Abstract: The subject of this dissertation belongs to scientific field of spectral graph theory, a young branch of mathematical combinatorics, i.e. graph theory, which finds important applications in many areas, such as chemistry, physics, computer science, telecommunications, sociology, etc., and various fields of mathematics. Spectral graph theory connects basic properties and the structure of a graph with characteristics of the spectra of its matrices (adjacency matrix, Laplacian matrix, etc.). In this dissertation we only work with the adjacency matrix. The second largest eigenvalue of the adjacency matrix of a graph (or, simply, second largest eigenvalue of a graph), as well as its distance from the largest eigenvalue, are very important especially in applications of spectral graph theory in computer science. The property of a graph that one of its eigenvalues does not exceed some given value is a hereditary one; therefore, many of the investigations of this kind have been directed at finding the maximal allowed graphs, or minimal forbidden graphs for that property. In this dissertation we determine some classes of graphs whose second largest eigenvalue does not exceed some given value, and, for that purpose, we develop some very useful tools. In methodological sense, investigations in this dissertation represent a combined approach consisting of application of the algebraic apparatus and methods of spectral graph theory and combinatorial reasoning, whilst at some stages the expert system newGRAPH has been used. The dissertation consists of eight chapters, each of which is divided into subchapters. In the beginning, some important previous work is shown, and afterwards we present some original elements of the algebraic and combinatorial apparatus that speed up and simplify the further work. We define some mappings between certain families of graphs, some of which preserve the sign of the expression 2 2 , and, using them, we describe and systematize some (already known) results in a new way. Further on we completely determine all maximal reflexive tricyclic cacti which are not RSdecidable and whose cycles do not form a bundle, from the classes 1 R and 3 R , and we give some partial results about the class 2 R , using previously induced mappings (until now only the graphs from the remaining class 4 R have been completely determined [40], [46]). Next, we completely describe all minimal forbidden graphs in the class of bicyclic graphs with a bridge, and all minimal forbidden graphs in the class 3 R  the approach that so far has never been used with reflexive graphs. Then we determine the maximal number of the cycles for RSundecidable reflexive cacti whose cycles do form a bundle, and, therefore, generally for RSundecidable reflexive cacti and we describe three classes of maximal reflexive RSundecidable reflexive cacti that contain a bundle. Further on, some of the previous results are generalized: the generalized RStheorem is stated and proved (socalled GRStheorem) for any r , r 0 ; previously induced mappings are generalized, their properties are proved and various examples of classes of graphs with the property 2 r (for r 0 ) are given. Based on this, we describe all GRSundecidable maximal graphs for the property 2 2 in the class of unicyclic and multicyclic graphs, and also all RSundecidable maximal θgraphs for this property as well as all GRSundecidable maximal trees with the property 2 5 1 2 . Furthermore, we investigate the limit 3 (as in [28]) and we describe all trees with the diameter 3 and the diameter larger than 8, with the property 2 3 , as well as all GRSundecidable multicyclic cacti with the same property. Finally, we introduce and apply socalled σmodifications of Smith trees. We describe seven σmodifications and corresponding extensions, and we notice the appearance in (already known) results in the class of multicyclic reflexive cacti with 4 cycles. Applying some extensions to certain families of tricyclic cacti, we obtained the results in the class of multicyclic reflexive cacti with 4 cycles, using a different approach [48]. Finally, in the conclusion, we suggest some possible directions of further investigations. URI: http://hdl.handle.net/123456789/4445 Files in this item: 1
Mihailovic_Bojana.pdf ( 6.960Mb ) 
Acketa, Dragan (Novi Sad , 1984)[more][less]