Browsing Doctoral Dissertations by Title
-
Kordić, Stevan (Beograd , 2016)[more][less]
Abstract: Constrain satisfaction problems including the optimisation problems are among the most important problems of discrete mathematics with wide area of application in mathematics itself and in the applied mathematics. Dissertation study optimisation problem and presents an original method for finding its exact solution. The name of the method is Sedimentation Algorithm, which is introduced together with two heuristics. It belongs to the class of branch-and-bound algorithms, which uses backtracking and forward checking techniques. The Sedimentation Algorithm is proven to be totally correct. Ability of the Sedimentation Algorithm to solve different type of problems is demonstrated in dissertation by its application on the Boolean satisfiability problems, the Whitehead Minimisation Problem and the Berth Allocation Problem in container port. The best results are obtained for Berth Allocation Problem, because its modelling for Sedimentation Algorithm includes all available optimisation techniques of the method. The precise complexity estimation of the Sedimentation Algorithm for the Berth Allocation Problem is established. Experimental results verify that the Sedimentation Algorithm is capable to solve the Berth Allocation Problem on the state of art level. URI: http://hdl.handle.net/123456789/4413 Files in this item: 1
StevanKordic.pdf ( 2.477Mb ) -
Kapetanović, Miodrag (Belgrade)[more][less]
-
Cvetković, Predrag (Belgrade , 1976)[more][less]
-
Arbutina, Bojan (Faculty of Mathematics, University of Belgrade , 2009)[more][less]
Abstract: The main research topic of this dissertation are extreme mass ratio contact close binary systems, q 0.1, of W Ursae Majoris (W UMa) type. These close binaries (CBs) represent an interesting class of objects in which ”normal”, approximately one solar mass main-sequence star is in contact with a significantly less massive companion, M2 ∼ 0.1 M . Earlier theoretical investigations of these systems found that there is a minimum mass ratio qmin = M2/M1 ≈ 0.085 − 0.095 (obtained for n = 3 polytrope - fully radiative primary) above which these CBs are stable and could be observed. If the mass ratio is lower than qmin, or, equivalently, if orbital angular momentum is only about three times larger than the spin angular momentum of a massive primary, a tidal instability develops (Darwin’s instability) forcing eventually the stars to merge into a single, rapidly rotating object (such as FK Com-type stars or blue stragglers). However, there appear to be some W UMa-type CBs with empirically obtained values for the mass ratio below the theoretical limit for stability. The aim of this dissertation is to try to resolve the discrepancy between theory and observations by considering rotating polytropes. By including in theory the effects of higher central condensation due to rotation we were able to reduce qmin to the new theoretical value qmin = 0.070 − 0.074, for the overcontact degree f = 0 − 1, which is more consistent with the observed population. Other candidate systems for stellar mergers such as AM CVn-type stars have also been discussed in the dissertation. URI: http://hdl.handle.net/123456789/716 Files in this item: 1
phdBojanArbutina.pdf ( 6.326Mb ) -
Ćelić, Momir (Banjaluka , 1986)[more][less]
-
Rakić, Nemanja (Beograd , 2022)[more][less]
Abstract: Active Galactic Nuclei (AGN) are galaxy cores in which a supermassive black hole is being surrounded with an accretion disk, which emits powerful continuum emission. This continuum ionizes the surrounding gas which than emits the broad lines, detected in the optical band. The aim of this thesis is to model complex optical spectra of large numbers of type 1 AGNs (AGNs with broad emission lines) and study physical properties of the emitting gas, so-called broad line region. Understanding the physics of the broad line region, as well as the physics of AGN as a whole is important for understanding the galaxy formation and evolution. This thesis investigate a large sample of type 1 AGNs taken from the SDSS spectral database. In total, over 1500 spectra were analyzed, for which a completely new software for modeling of optical AGN spectra (named FANTASY) was developed. Automated simultaneous modeling of all emission components in the optical spectrum enables reliable measurement of spectral parameters, which can have significant application in future investigation of large AGN samples, collected within the upcoming large spectroscopic sky surveys. The physical properties in the broad line region were studied through the analysis of a still unexplained (anti)correlation between the equivalent width of the emission lines and the conti- nuum flux (the so-called Baldwin effect). It has been confirmed that the intrinsic Baldwin effect exists in the Balmer hydrogen lines in type 1 AGNs, and that a potential physical explanati- on for this effect might be presence of an additional non-ionizing optical continuum emission, which originates from the broad line region. Finally, the significant outcome of this thesis is is the new publicly available software package FANTASY, which can be widely used in the future spectral analysis of AGN. URI: http://hdl.handle.net/123456789/5451 Files in this item: 1
Doktorska_disertacija_nrakic.pdf ( 5.806Mb ) -
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 three-dimensional hydrodynamic modelling and non-linear 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 non-resonant 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 Sedov-Taylor 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 three-dimensional hydrodynamic modelling and non-linear 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 non-resonant 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 Sedov-Taylor 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 experi-mental evaluation on relevant benchmarks shows thatSPECSis efficient and comparing tostate-of-the-art solvers, it gives more precise results in a shorter amount of time, while suppor-ting 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 first-order 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 well-known 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 NP-hard 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: 2-swap, 2-swap 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 col-lection 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 implemen-tation 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 semi-automatic 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 semi-automatic 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 rule-based 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 (clas-sification 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 rule-based 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 rule-based 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 ) -
Džamić, Dušan (Beograd , 2021)[more][less]
Abstract: The theory of complex networks has proven to be very important inthe study of the characteristics and structure of various complex systems. In thelast two decades, a large number of researches have been directed towards thedevelopment of methods for clustering in complex networks. In 2012, Center forDiscrete Mathematics and Theoretical Computer Science (DIMACS), which is awell-known consortium of prestigious academic institutions (Rutgers University,Princeton, Colombia) and research laboratories (Microsoft, IBM, AT & T, NEC),included the problem of clustering in complex networks on the list of the mostimportant problems and challenges in computer science. Clustering in complexnetworks can be applied in a variety of contexts to achieve different goals, andtherefore, there is no generally accepted definition of a cluster. For this reason,different approaches are used in developing clustering methods. An approach thathas attracted the most attention of researchers involves two subproblems: defininga function to determine the quality of a partition and constructing methods to finda partition that has the maximum value of the defined quality function. In thisapproach, the problem of clustering is formulated as the problem of combinatorialoptimization and various methods of mathematical optimization can be used tosolve it. One of the most commonly used quality function is the modularity.Clustering by modularity maximization, i.e., finding a partition with the max-imum value of modularity, is NP-hard problem. Thus, only heuristic algorithmsare suitable of processing large datasets. In this dissertation, a novel method formodularity maximization based on the variable neighborhood search heuristic isproposed. For the purpose of efficient application in large-scale complex networks,a procedure for decomposition of the problem into smaller subproblems is devel-oped. In addition, a mechanism for overcoming local maxima of modularity isimproved using criteria for occasional acceptance of solution which is worse thanthe current one. DIMACS instances are used to test the proposed method, and theobtained results are compared with the best ones presented in the literature, ob-tained by two methods developed in DIMACS implementation challenge in 2012.In addition, the obtained results are compared with the results of six methodsdeveloped after 2012, from the literature. A comparative analysis of the resultsshows that the proposed method outperforms the existing methods for modularitymaximization and improves the best known solutions on 9 out of 13 hard instances. Clustering by modularity maximization is not suitable for detecting small clus-ters in large networks. For this reason, a new function for measuring the qualityof a partition has been proposed in the dissertation. Through three theorems, itis shown that the new measure, called E-function, has the potential to identifyclusters in the network and overcome limits of modularity. For testing the pro-posed E-function and comparison with the modularity function, a generic variableneighborhood method is developed to optimize the considered quality function.Computational experiments are performed on generated and real instances fromthe literature for which the correct division into clusters is known. The resultsshow that the expected clusters can be identified, both on artificial and real in-stances, by maximizing the E-function. URI: http://hdl.handle.net/123456789/5214 Files in this item: 1
teza_dzamic.pdf ( 3.074Mb ) -
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 )