Browsing Doctoral Dissertations by Title
-
Mosurović, Milenko (Belgrade)[more][less]
Abstract: The results of this thesis are connected with papers of Wolter and Zakharyashchev in which various expressive and decidable description logics with epistemic, temporal, and dynamic operators are constructed, but the complexity of the satisfaction problem in these logics has remained unclear. The thesis consists of seven chapters. Chapter 1 and 5 contain the notions and properties which are used in other chapters. In Chapter 2 a fragment of the logic DIF, which is called D_1IF , is considered and new constructions and proofs for that logic are given. In Chapter 3 a fragment of the logic DIO, which is called D_1IO, is studied. In Chapter 4 by using the results from Chapter 2 and Chapter 3 NEXPTIME algorithm, which tests whether a structure simple quasi-world or not, is constructed and it is used in Chapter 7. In Chapters 6 and 7 it is shown that the modal description logics of Wolter and Zakharyashchev based on arbitrary frames are NEXPTIME-complete, no matter whether the underlying description logic is ALC, CI or C_1IQ. Moreover, it is shown that these logics based on S5-frames (for knowledge), and KD45-frames (for beliefs) are also NEXPTIME-complete. Finally, the following property is shown: the description logics of Wolter and Zakharyashchev based on N-frames (for time) are EXPSPACE-complete, no matter whether the underlying description logic is ALC_U, CI_U or C_1IQ_U. URI: http://hdl.handle.net/123456789/192 Files in this item: 1
phdMilenkoMosurovic.pdf ( 3.024Mb ) -
Kanan, Asmaa (Beograd , 2013)[more][less]
Abstract: Semiring with zero and identity is an algebraic structure which generalizes a ring. Namely, while a ring under addition is a group, a semiring is only a monoid. The lack of substraction makes this structure far more difficult for investigation than a ring. The subject of investigation in this thesis are matrices over commutative semirings (wiht zero and identity). Motivation for this study is contained in an attempt to determine which properties for matrices over commutative rings may be extended to matrices over commutative semirings, and, also, which is closely connected to this question, how can the properties of modules over rings be extended to semimodules over semirings. One may distinguish three types of the obtained results. First, the known results concerning dimension of spaces of n-tuples of elements from a semiring are extended to a new class of semirings from the known ones until now, and some errors from a paper by other authors are corrected. This question is closely related to the question of invertibility of matrices over semirings. Second type of results concerns investigation of zero divisors in a semiring of all matrices over commutative semirings, in particular for a class of inverse semirings (which are those semirings for which there exists some sort of a generalized inverse with respect to addition). Because of the lack of substraction, one cannot use the determinant, as in the case of matrices over commutative semirings, but, because of the fact that the semirings in question are inverse semirings, it is possible to define some sort of determinant in this case, which allows the formulation of corresponding results in this case. It is interesting that for a class of matrices for which the results are obtained, left and right zero divisors may differ, which is not the case for commutative rings. The third type of results is about the question of introducing a new rank for matrices over commutative semirings. For such matrices, there already exists a number of rank functions, generalizing the rank function for matrices over fields. In this thesis, a new rank function is proposed, which is based on the permanent, which is possible to define for semirings, unlike the determinant, and which has good enough properties to allow a definition of rank in such a way. URI: http://hdl.handle.net/123456789/4274 Files in this item: 1
phdKanan_Asmaa.pdf ( 1.221Mb ) -
Dedagić, Fehim (Priština)[more][less]
-
Karapetrović, Boban (Beograd , 2017)[more][less]
Abstract: In this thesis, we study the in nite Hilbert matrix viewed as an operator, called the Hilbert matrix operator and denoted by H and Libera operator, denoted by L, on the classical spaces of holomorphic functions on the unit disk in the complex plane. It is well known that the Hilbert matrix operator H is a bounded operator from the Bergman space Ap into Ap if and only if 2 < p < 1. Also, it is known that the norm of the Hilbert matrix operator H on the Bergman space Ap is equal sin 2 p , when 4 p < 1, and it was conjectured that kHkAp!Ap = sin 2 p ; when 2 < p < 4. In this thesis we prove this conjecture. We nd the lower bound for the norm of the Hilbert matrix operator H on the weighted Bergman space Ap; , kHkAp; !Ap; sin ( +2) p ; for 1 < + 2 < p: We show that if 4 2( + 2) p, then kHkAp; !Ap; = sin ( +2) p ; while in the case 2 +2 < p < 2( +2), upper bound for the norm kHkAp; !Ap; , better then known, is obtained. We prove that the Hilbert matrix operator H is bounded on the Besov spaces Hp;q; if and only if 0 < p; ; = 1 p + 1 < 1. In particular, operator H is bounded on the Bergman space Ap; if and only if 1 < + 2 < p and it is bounded on the Dirichlet space Dp = Ap; 1 if and only if maxf1; p 2g < < 2p 2. We also show that if > 2 and 0 < " 2, then the logarithmically weighted Bergman space A2 log is mapped by the Hilbert matrix operator H into the space A2 log 2" . If 2 R, then the Hilbert matrix operator H maps logarithmically weighted Bloch space Blog into Blog +1. We also prove that operator H maps logarithmically weighted Hardy-Bloch space B1 log , when 0, into B1 log 1 and that this result is sharp. Also, we have that the space VMOA is not mapped by the Hilbert matrix operator H into the Bloch space B. On the other hand, we nd that the Libera operator L is bounded on the Besov space Hp;q; if and only if 0 < p; ; = 1 p + 1. Then, we prove that if > 1, then the logarithmically weighted Bergman space A2 log is mapped by the Libera operator L into the space A2 log 1 , while if 2 R, then the Libera operator L maps logarithmically weighted Bloch space Blog into itself. If > 0, we have that operator L maps logarithmically weighted Hardy-Bloch space B1 log into B1 log 1 and we show that this result is sharp. The well known conjecture due to Korenblum about maximum principle in Bergman space Ap states: Let 0 < p < 1. Then there exists a constant 0 < c < 1 with the following property. If f and g are holomorphic functions in the unit disk D, such that jf(z)j jg(z)j for all c < jzj < 1, then kfkAp kgkAp . Hayman proved Korenblum's conjecture for p = 2 and Hinkkanen generalized this result, by proving conjecture for all 1 p < 1. The case 0 < p < 1 of conjecture still remains open. In this thesis we resolve this case of the Korenblum's conjecture, by proving that Korenblum's maximum principle in Bergman space Ap does not hold when 0 < p < 1. URI: http://hdl.handle.net/123456789/4497 Files in this item: 1
Boban_Karapetrovic_doktorska_disertacija.pdf ( 1.392Mb ) -
Filipović, Vladimir (Beograd , 2006)[more][less]
URI: http://hdl.handle.net/123456789/2478 Files in this item: 1
phdFilipovicVladimir.pdf ( 2.145Mb ) -
Ivković, Zoran (Belgrade , 1964)[more][less]
-
Kalajdžić, Gojko (Belgrade , 1982)[more][less]
-
Makragić, Milica (Beograd , 2018)[more][less]
Abstract: This doctoral dissertation comprises two parts. Trigonometric polynomial rings are the central topic of the first part of the dissertation. It is presented that the ring of complex trigonometric polynomials, C [cos x, sin x ], is a unique factorization domain, and that the ring of real trigonometric polynomials, R [cos x, sin x ], is not a unique factorization domain. Necessary and sufficient conditions for the case when in the ring C [cos x, sin x ], unlike the ring R [cos x, sin x ], the degree of the product of two trigonometric polynomials is not equal to the sum of degrees of its factors, are given. The theory of trigonometric polynomials is extended to hyperbolic-trigonometric polynomials, or HT-polynomials for short, which are defined similarly to trigonome- tric polynomials. Real or complex HT-polynomials form a ring and even an integral domain R [cosh x, sinh x ], or C [cosh x, sinh x ]. Factorization in these domains is con- sidered, and it is shown that these are unique factorization domains. The irreducible elements, as well as the form of the maximal ideals of both these domains are deter- mined. The algorithms for dividing, factoring, computing greatest common divisors, as well as the algorithms for simplifying ratios of two HT-polynomials are considered over the field of rational numbers. In the second part of the dissertation, related to applications, two methods of proving inequalities of the form f ( x ) > 0 are described over the given finite in- terval ( a,b ) ⊂ R , a ≤ 0 ≤ b , which by using the finite Maclaurin series expan- sion generate polynomial approximations, when the function f ( x ) is element of the ring extension of R [cos x, sin x ], or R [cosh x, sinh x ], denoted by R [ x, cos x, sin x ], or R [ x, cosh x, sinh x ]. The completeness of the presented methods is proved and the concrete results of these methods are illustrated through examples of proving actual inequalities. URI: http://hdl.handle.net/123456789/4745 Files in this item: 1
Disertacija_Milica_Makragic.pdf ( 2.169Mb ) -
Bajšanski, Bogdan (Novi Sad)[more][less]
-
Marković, Sima (Belgrade , 1912)[more][less]
-
Ajajbegović, Jusuf (Sarajevo)[more][less]
URI: http://hdl.handle.net/123456789/156 Files in this item: 1
phdJusufAlajbegovicpdf.pdf ( 3.012Mb ) -
Erić, Lj. Aleksandra (Belgrade , 2012)[more][less]
Abstract: This thesis concerns P-vertices and P-set of non-singular acyclic matrices A and also singular acyclic matrices. It was shown that each singular matrix of order n has at most n ¡ 2 P-vertices. Also, it is shown that this does not hold for non-singular acyclic matrices by constructing non-singular acyclic matrices whose graphs are T having n¡1 ( or n) P-vertices. These matrices also achieve maximum size of P-set over non-singular acyclic matrices whose graphs are T. In this thesis, there is classi¯cation of the trees for which there is non- singular matrix where each vertex is P-vertex. In particular, it is shown that such trees have an even number of vertices. Both results provide answer to questions proposed by I.-J. Kim and B. L. Shader. In the end, related classi¯cations on non-singular trees with the size of a P-set bounded are addressed. Also, it is shown that double star DSn with n vertices, is an example of a tree such that, for each non-singular matrix A whose graph is DSn the number of P-vertices of A is less than n¡2. This example provides a positive answer to a question proposed recently by Kim and Shader. A recent classi¯cation of those trees for which each of associated acyclic matrices has distinct eigenvalues whenever the diagonal entries are distinct was established. Here is analyze of maximum number of distinct diagonal entries, and corresponding location, in order to preserve that multiplicity characterization. Recently, the multiplicities of eigenvalues of ©-binary tree was analyzed. This paper carry this discussion forward extending their results to larger family of trees, namely, the wide double path, a tree consisting of two paths that are joined by another path. Some introductory considerations for dumbbell graphs are mentioned re- garding the maximum multiplicity of the eigenvalues. URI: http://hdl.handle.net/123456789/2488 Files in this item: 1
Eric_Aleksandra.pdf ( 2.125Mb ) -
Vučetić, Milica (Beograd , 2017)[more][less]
Abstract: In this dissertation we discuss in uence of H emission from supernova remnants (SNRs) on H derived star formation rates (SFRs). We present the detection of 16 optical SNR candidates in a part of nearby spiral galaxy IC342, and two more poten- tial SNRs in NGC 185 galaxy. The candidates were detected by applying [S ii]/H ratio criterion on observations made with the 2 m telescope at Rozhen National Astronomical Observatory in Bulgaria. Also, in this dissertation we present the most up-to-date list of nearby galaxies with optically detected SNRs. When deri- ving H star formation rates, assumption is that the radiation is coming from the ionized gas surrounding hot and young stars { Hii regions. In this case, emission from SNRs contaminates H ux used to derive SFR. We found that the contribu- tion of SNRs' ux to the total H ux, for 18 galaxies in our sample of galaxies with optical SNRs, is 5 5%. Due to the observational selection effects, the SNR contamination of SFRs derived herein represents only a lower limit. M83 is the galaxy with the most detected optical SNRs and in this galaxy SNRs account for 9 per cent of the total H emission. We expect that fraction similar to this would be close to the real contribution of SNR emission to the total H emission in spiral galaxies. Using two dwarf galaxies as an example, Holmberg IX and NGC 185, we show that the contamination of H ux by SNRs, or other sources, can be much larger, up to ten times higher than total H ux coming from Hii regions of the observed galaxy. URI: http://hdl.handle.net/123456789/4501 Files in this item: 1
Milica_disertacija_21.03.2017.pdf ( 26.28Mb ) -
Vučetić, Milica (Beograd , 2017)[more][less]
Abstract: In this dissertation we discuss in uence of H emission from supernova remnants (SNRs) on H derived star formation rates (SFRs). We present the detection of 16 optical SNR candidates in a part of nearby spiral galaxy IC342, and two more poten- tial SNRs in NGC 185 galaxy. The candidates were detected by applying [S ii]/H ratio criterion on observations made with the 2 m telescope at Rozhen National Astronomical Observatory in Bulgaria. Also, in this dissertation we present the most up-to-date list of nearby galaxies with optically detected SNRs. When deri- ving H star formation rates, assumption is that the radiation is coming from the ionized gas surrounding hot and young stars { Hii regions. In this case, emission from SNRs contaminates H ux used to derive SFR. We found that the contribu- tion of SNRs' ux to the total H ux, for 18 galaxies in our sample of galaxies with optical SNRs, is 5 5%. Due to the observational selection effects, the SNR contamination of SFRs derived herein represents only a lower limit. M83 is the galaxy with the most detected optical SNRs and in this galaxy SNRs account for 9 per cent of the total H emission. We expect that fraction similar to this would be close to the real contribution of SNR emission to the total H emission in spiral galaxies. Using two dwarf galaxies as an example, Holmberg IX and NGC 185, we show that the contamination of H ux by SNRs, or other sources, can be much larger, up to ten times higher than total H ux coming from Hii regions of the observed galaxy. URI: http://hdl.handle.net/123456789/4507 Files in this item: 1
Milica_disertacija_21.03.2017.pdf ( 26.28Mb ) -
Vučković, Vesna (Univerzitet u Beogradu – Matematički fakultet , 2010)[more][less]
Abstract: This dissertation deals with digital watermark for grayscale images. Digital watermark in copyright protection needs to satisfy detectability and fidelity conditions. If watermark is embedded stronger, it will be likely detectable, but it will be noticeable too, and it will jeopardize host image quality. Here, for a well known, AWGN watermark, optimal embedding strength (minimal which guaranties watermark detectability) is determined. Optimal strength is analyzed - for effective embedding (watermark needs to be detectable immediately after embedding) - for watermark to be robust against expected modification (it needs to be detectable in image which will be after embedding subjected to this modification) For effective AWGN watermark embedding, mathematical formula for optimal strength calculating is derived. For watermark robust against expected modification, one algorithm for optimal strength is given. Among all image modifications (valumetric and geometric), lossy compression surely has an important place. Images which nowadays we can find on Net are mostly in some lossy compressed form. In such circumstances, embedded watermark will also surely be exposed to lossy compression. This is why in this dissertation particular attention is devoted to lossy compression. Thus, its considerable part deals with AWGN watermark optimal strength, in spatial and in transform domains. Specially, here is analyzed embedding in some image subchannels in the block DCT domain, and quantization noise impact on the embedded watermark message. URI: http://hdl.handle.net/123456789/1073 Files in this item: 1
PhDVVuckovic.pdf ( 4.249Mb ) -
Banjević, Dragan (Belgrade)[more][less]
-
Šestak, Ivan (Belgrade , 1987)[more][less]
-
Vuković, Josif (Belgrade , 1984)[more][less]
-
Glišović, Nataša (Beograd , 2018)[more][less]
Abstract: In this doctoral dissertation the modelling process has been taken into consideration in the presence of uncertainty. Two types of problems were analyzed: one is the optimization of the benefit/costs tradeoff during the distribution of the projects and the other is the classification of data described by the attributes among which some are missing. The basic problems during the modelling of the decision making in the presence of uncertainty are the choice of the adequate treatment of uncertainty and the choice of the method for making a decision. One of the aims of the work is investigating the benefits of applying the metaheuristic algorithms on the considered optimization problems. The main measure for the evaluation of their performances is the value of objective function (for both problems: optimization of benefit/costs tradeoff during the project scheduling and clustering of incomplete data). Considering the project scheduling problem the level of satisfaction related to the problem constraints could also be taken into account. The other evaluation criteria of the applied metaheuristic methods is the time required for finding the solution. The influence of the parameters which control the algorithms of the metaheuristic methods is examined, as well as their appropriate values leading to the maximum performances of the implementation could be reached on the tested examples of the considered problems. As for the optimization problem of the profit/costs tradeoff, the uncertainty is modelled by applying the triangle fuzzy problems and then the metaheuristic methods, simulated annealing and genetic algorithm were applied for solving the obtained fuzzy optimization problem. The tested problems are formulated by the fuzzification method which was suggested by (Ribeiro et al. 1999). The represented experimental results for the set of fuzzy problems show the efficiency of the applied methods: simulated annealing and genetic algorithm. Genetic algorithm seems to produce slightly better solution than the simulated annealing. However, both methods out performed the existing form the literature for about 20%. The secund part of the work deals with the clustering data problem with the missing values of the attributes and making decisions in such circumstances. The main phases in solving the considered problem are finding the most appropriate distance, which will be used in the cases when the data are missing for some reasons and choosing the method for solving the clustering problem. As the theoretical and practical contribution, the metric, based on the logic principles, was proposed. By applying the probability, the theorem was proved defining the values of the weighting coefficients related to attributes that describe the objects for clustering. The proposed metric was implemented in the variable neighborhood search metaheuristic method as well as in some of its modifications. The implemented methods have been applied on the real life problems from the literature. Classifying the patients who suffer from some auto-immune diseases, stored in the database of Clinical Centre of Serbia, the precision of the clustering of 93.33% was achieved. As another real life example, seven databases of the European Commision (Board), which contain the data for the mail service, have been analyzed. The clustering efficiency of 90% - 96.96% was achieved. In order to compare the efficiency of the approach based on the variable neighborhood search method, nine databases available on the internet were used and the obtained results were compared with the existing ones from the literature. The experiments showed large stability of variable neighborhood search method: in eight out of nine cases the best solution was reached in all hundred repetitions. Besides that, the quality of the obtained solutons have considerably surpassed the results from the literature. URI: http://hdl.handle.net/123456789/4710 Files in this item: 1
Nglis_DoktorskaDisertacija.pdf ( 3.362Mb ) -
Jovović, Ivana (Belgrade , 2013)[more][less]
Abstract: This dissertation deals with an application of some linear algebra techniques for solving problems of reduction of system of linear operator equation of the form A(x1) = b11x1 + b12x2 + : : : + b1nxn + '1 A(x2) = b21x1 + b22x2 + : : : + b2nxn + '2 ... A(xn) = bn1x1 + bn2x2 + : : : + bnnxn + 'n; where B = [bij ]n n is matrix over the eld K, A is linear operator on the vector space V over K and where '1; '2; : : : ; 'n are vectors in V . In particular, we consider reduction of such system under the action of the general linear group GL(n;K) and also reduction by using the characteristic polynomial B( ) of the matrix B and recurrence for the coe cients of the adjugate matrix of the characteristic matrix I B of the matrix B. The idea is to use rational and Jordan canonical forms to reduce the linear system of operator equations to an equivalent partially reduced system, i.e. to decompose the initial system into several uncoupled systems. This represents a new application of doubly companion matrix introduced by J.C. Butcher in [5]. In this work we are also concerned with transformation of the linear system of operator equation into totally reduced system, i. e. completely decoupled system of higher order linear operator equations. This results are related to results given by T. Downs in [13]. The thesis consists of two parts. The rst part deals with properties of rational and Jordan canonical form. We start with Fundamental Theorem of Finitely Generated Modules Over a Principle Ideal Domain. If we consider nite dimensional vector space V over K as module over the ring K[x] of polynomials in x with coe cients in K, the Fundamental Theorem implies that there is a basis for V so that the associated matrix for B is in rational or Jordan form. The rst section is adapted from Abstract Algebra of D. S. Dummit i R. M. Foote [14]. In the second section we look more closely at Hermite, Smith, rational and Jordan form and establish the relation between them. The structure of the similarity transformation matrix is also described. Some of theorem are considered from several aspects. This section provides a detailed exposition of normal forms using [14, 19, 37, 34, 35, 20, 57] and [1, 22, 48, 52, 54, 60]. The second part concerns with author's original contribution and it relies on papers [42, 43]. First we illustrate methods of the partial and the total reduction of systems v in two or three unknowns and then we study reductions of systems in n unknowns. The partial reduction requests changing of basis so that the system matrix is in the rational or Jordan form. We also treat the total reduction of the obtained partially reduced systems in this manner. Subsection 4.4. "Total Reduction for Linear Systems of Operator Equations with System Matrix in Companion Form" is one interesting way to proceed consideration started in previously mentioned works. It is based on papers of L. Brand [3, 4]. The fth section is generalization of the forth. Here we examine systems in n unknowns and with di erent linear operators. We introduce the notion of characteristic polynomial in more than one unknown - generalized characteristic polynomial and a method for total reduction by nding adjugate matrix of the generalized characteristic matrix of the system matrix. The sixth section is a summary of applications and examples of methods for partial and total reduction. There are some examples of the rst and higher order linear systems of di erential equations and di erent approaches for calculating rational and Jordan canonical forms. The last section is devoted to the study of di erential transcendence of the solution of the rst order linear system of di erential equations with complex coe cients, where exactly one of the following meromorphic functions '1; '2; : : : ; 'n is di erentially transcendental, using method of total reduction. We review some of the standard facts on di erential transcendence following books [45, 44, 39, 50, 7, 33, 15, 23, 36]. URI: http://hdl.handle.net/123456789/2585 Files in this item: 1
Jovović_Ivana.pdf ( 2.063Mb )