University of Patras
Πιθανοτική μελέτη του τμήματος συγκέντρωσης ροών σε δυαδικές ακολουθίες
Abstract
dc:descriptionWe consider a finite sequence of binary (failure - success) random variables (RVs). The sequence's elements (RVs) are assumed to be independent (identical/nonidentical) or dependent (exchangeable or homogeneous/nonhomogeneous Markov dependent). The aim of the present Ph.D. thesis is the probabilistic study of the minimum sequence's segment in which all runs of successes with length greater than or equal to a fixed size (i.e. a threashold length) are concentrated. In order to probabilisticaly locate and describe the concentration segment, we define run statistics (RVs) relative to it. The statistics represent the number of success runs, the number of successes in the success runs, the length of the longest success run, and the length of the segment and the position (starting/ending) of it in the sequence. The first four RVs are univariate and the fifth is a bivariate one. In addition, we define a trivariate RV having as its components the counting successes and runs of successes RVs and the length of the concentration segment. In the thesis, the conditional probability mass functions of the last three RVs are established, for the first time, given that the number of success runs in the sequence is at least two. This treatment involves all the prementioned RVs. Exact expressions for probabilities and probability functions are obtained in the form of recursive schemes and sums involving binomial coefficients. In deriving the expressions we take into account the particular characteristics of the internal structure of the examined binary sequences as well as the definition of the RVs on such sequences. The thesis contains six chapters which are classified into three parts (A, B and Γ). Each part consists of two chapters having a common feature. More specifically, the thesis is organized as follows. Part A comprises Chapters 1 and 2. Chapter 1 mainly concerns with preliminary concepts and definitions. They refer to probability models (random sources), along with their interrelations, appropriate for describing the internal structure of a sequence of independent or dependent binary RVs. Statistics, relevant to the subject matter of the thesis, are formally defined in Chapter 2. A motivation for their usefulness is also discussed. As it was mentioned above, the statistics enumerate successes and success runs in the sequence, denote the length and the position of the sequence's segment containing runs of successes and they jointly describe the number of success runs, the number of successes in these runs and the length of the concentration segment of the success runs in the sequence. The probabilistic treatment of the statistics presented in Chapter 2 and defined on binary sequences discussed in Chapter 1 is the subject of Part B comprising Chapters 3 and 4. In Chapter 3, the number of success runs of length greater than or equal to a fixed length is examined. Chapter 4 deals with the segment of the concentration of such success runs in the sequence. More specifically, in Sections 4.1, 4.2 and 4.3, we obtain results referring to the length of the concentration segment, the position of this segment in the sequence and how the number of success runs, the number of successes in these runs and the length of their concentration segment in the sequence are jointly distributed, respectively. The theoretical results are further illustrated via indicative numerical examples. The applicability of the material developed in Parts A and B is clarified in Part Γ. The latter part comprises Chapters 5 and 6. Chapter 5 provides a detailed numerical study, accompanied by a discussion on its findings, of the probabilities and probability functions derived in Part B. Potential applications along with extensive and illustrative numerics are presented in Chapter 6. The vast majority of the thesis' results are new. Part of these is published (is under publication) in the papers of Makri et al. (2015) and Arapis et al. (2016a, 2017a, 2017b).
Degree
thesis:*- Grantor dc:publisher
- University of Patras
- Year dc:date
- 2017
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Arapis, Anastasios
- Αράπης, Αναστάσιος
Subjects
dc:subject × 18Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/42889