The Generalised Star-Height Problem is an open question in the field of formal language theory that concerns a measure of complexity on the class of regular languages; specifically, it asks whether or not there exists an algorithm to determine the generalised star-height of a given regular language. Rather surprisingly, it is not yet known whether there exists a regular language of generalised star-height greater than one. Motivated by a theorem of Thérien, we first take a combinatorial approach to the problem and consider the languages in which every word features a fixed contiguous subword an exact number of times. We show that these languages are all of generalised star-height zero. Similarly, we consider the languages in which every word features a fixed contiguous subword a prescribed number of times modulo a fixed number and show that these languages are all of generalised star-height at most one. Using these combinatorial results, we initiate work on identifying the generalised star-height of the languages that are recognised by finite semigroups. To do this, we establish the generalised star-height of languages recognised by Rees zero-matrix semigroups over nilpotent groups of classes zero and one before considering Rees zero-matrix semigroups over monogenic semigroups. Finally, we explore the generalised star-height of languages recognised by finite groups of a given order. We do this through the use of finite state automata and 'count arrows' to examine semidirect products of the form A x Zr where A is an abelian group and Zr is the cyclic group of order r.

The syntactic complexity of a regular language is the cardinality of its syntactic semigroup. The syntactic complexity of a subclass of the class of regular languages is the maximal syntactic complexity of languages in that class, taken as a function of the state complexity n of these languages. We study the syntactic complexity of suffix-, bifix-, and factor-free regular languages, star-free languages including three subclasses, and R- and J-trivial regular languages. We found upper bounds on the syntactic complexities of these classes of languages. For R- and J-trivial regular languages, the upper bounds are n! and ⌊e(n-1)!⌋, respectively, and they are tight for n >= 1. Let C^n_k be the binomial coefficient ``n choose k''. For monotonic languages, the tight upper bound is C^{2n-1}_n. We also found tight upper bounds for partially monotonic and nearly monotonic languages. For the other classes of languages, we found tight upper bounds for languages with small state complexities, and we exhibited languages with maximal known syntactic complexities. We conjecture these lower bounds to be tight upper bounds for these languages. We also observed that, for some subclasses C of regular languages, the upper bound on state complexity of the reversal operation on languages in C can be met by languages in C with maximal syntactic complexity. For R- and J-trivial regular languages, we also determined tight upper bounds on the state complexity of the reversal operation.

The syntactic complexity of a regular language is the cardinality of its syntactic semigroup. The syntactic complexity of a subclass of the class of regular languages is the maximal syntactic complexity of languages in that class, taken as a function of the state complexity n of these languages. We study the syntactic complexity of suffix-, bifix-, and factor-free regular languages, star-free languages including three subclasses, and R- and J-trivial regular languages. We found upper bounds on the syntactic complexities of these classes of languages. For R- and J-trivial regular languages, the upper bounds are n! and ⌊e(n-1)!⌋, respectively, and they are tight for n >= 1. Let C^n_k be the binomial coefficient ``n choose k''. For monotonic languages, the tight upper bound is C^{2n-1}_n. We also found tight upper bounds for partially monotonic and nearly monotonic languages. For the other classes of languages, we found tight upper bounds for languages with small state complexities, and we exhibited languages with maximal known syntactic complexities. We conjecture these lower bounds to be tight upper bounds for these languages. We also observed that, for some subclasses C of regular languages, the upper bound on state complexity of the reversal operation on languages in C can be met by languages in C with maximal syntactic complexity. For R- and J-trivial regular languages, we also determined tight upper bounds on the state complexity of the reversal operation.

京都大学 / 0048 / 新制・課程博士 / 博士(理学) / 甲第18041号 / 理博第3919号 / 新制||理||1566(附属図書館) / 30899 / 京都大学大学院理学研究科数学・数理解析専攻 / (主査)准教授 西村 進, 教授 加藤 毅, 教授 長谷川 真人 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DGAM

Cette thèse présente des résultats obtenus dans deux domaines : la théorie des langages, et la théorie des graphes. En théorie des langages, on s’intéresse à des problèmes de caractérisation de classes de langages réguliers. Le problème générique consiste à déterminer si un langage régulier donné peut être défini dans un certain formalisme. Les méthodes actuelles font intervenir un problème plus général appelé séparation. On présente ici deux types de contributions : une généralisation d’un résultat de décidabilité au cadre des langages de mots infinis, ainsi que des bornes inférieures pour la complexité du problème de séparation. En théorie des graphes, on considère le problème classique de coloration de graphes, où on cherche à attribuer des couleurs aux sommets d’un graphe de sorte que les sommets adjacents reçoivent des couleurs différentes, le but étant d’utiliser le moins de couleurs possible. Dans le cas des graphes peu denses, la méthode de déchargement est un atout majeur. Elle a notamment joué un rôle décisif dans la preuve du théorème des quatre couleurs. Cette méthode peut être vue comme une construction non conventionnelle d’un schéma de preuve par induction, spécifique à la classe de graphes et à la propriété considérées, et où la validité du schéma est rarement immédiate. On utilise des variantes de la méthode de déchargement pour étudier deux types de problèmes de coloration. / In this thesis, we present results obtained in two fields: formal language theory and graph theory. In formal language theory, we consider some problems of characterization of classes of regular languages. The generic problem consists in determining whether a given regular language can be defined in a fixed formalism. The current approaches use a more general problem called separation. We present here two types of contributions: a generalization of a decidability result to the setting of infinite words, together with lower bounds for the complexity of the separation problem. In graph theory, we consider the classical problem of graph coloring, where we assign colors to vertices of a graph in such a way that two adjacent vertices receive different colors. The goal is to use the fewest colors. When the graphs are sparse, a crucial tool for this is the discharging method. It is most notably decisive in the proof of the Four-Color Theorem. This method can be seen as an unconventional construction of an inductive proof scheme, specific to the considered problem and graph class, where arguing the validity of the scheme is rarely immediate. We use variants of the discharging method to study two types of coloring problems.

Made available in DSpace on 2016-03-15T19:37:45Z (GMT). No. of bitstreams: 1 Wander Lairson Costa.pdf: 1957133 bytes, checksum: 6819580d97bb5eaca5ea04352fcda0b8 (MD5) Previous issue date: 2013-03-15 / Universidade Presbiteriana Mackenzie / Cellular automata are totally discrete systems that act locally in a simple and deterministic way, but whose resulting global behavior can be extremely complex. The set of possible global configurations in one finite time step for a CA can be described by a regular language, which in turn can be represented by a finite automaton, more precisely the so-called process graph, in which all states are initial and final. Here, we study the temporal evolution complexity of the elementary cellular automata (i.e., one-dimensional, binary, with radius 1), and related previous works are revisited and discussed, indicating problems and their consequences. We also start up a novel approach for the problem, substituting the process graph based representation that describes the configuration at each time step by adjacency matrices derived from them. In fact, we extend the classical adjacency matrix notation, as they cannot fully represent process graphs. With this new notation, we show that it is possible to obtain the algorithm to generate a process graph for an arbitrary finite time step for each of the rules at study. In conclusion, although advancing the limit graph problem, it still remains open, and we provide suggestions for further research. / Autômatos celulares são sistemas totalmente discretos que agem localmente de forma simples e determinística, mas cujo comportamento global resultante pode ser extremamente complexo. O conjunto de possíveis configurações globais em um passo de tempo t finito para um autômato celular pode ser descrito por uma linguagem regular, a qual por sua vez pode ser representada por meio de um autômato finito, mais precisamente, pelo chamado grafo de processo, em que todos os estados são iniciais e finais. Estuda-se aqui a complexidade da evolução temporal dos autômatos celulares elementares (i.e., unidimensionais, binários, de raio 1), e trabalhos anteriores são revisitados e discutidos, no quais apontam-se problemas e suas consequências. Também inicia-se uma nova abordagem para o problema, substituindo a representação dos grafos de processo que descrevem a configuração a cada passo de tempo por matrizes de adjacência deles derivadas. De fato, estende-se a notação clássica de matriz de adjacência, já que ela se mostra insuficiente para descrever completamente os grafos de processo em questão. Com essa nova notação, mostra-se que é possível obter o algoritmo que gere o grafo de processo de tempo t para cada uma das regras estudadas. Conclui-se que, embora houve avanços para o problema do grafo limite, este ainda permanece aberto, e sugestões para continuação da pesquisa são dadas.

The thesis aims to develop a method for extending the language platform Storm with a Java frontend.The project was conducted using an Action Research methodology and highlights triumphs andchallenges. Despite the significant overhead related to note generation and problem statementformulation, this methodology proved beneficial in identifying problems and providing the framework tosolve them. The first research question (RQ.1) evaluates to what extent the language platform Storm is suitable forimplementing the object oriented language Java. Using Storm, only a BNF and a specification for three-address code instructions are needed. Despite encountering difficulties during the implementation, theplatform offers tools that allow comprehensive customization of the new language's intended behaviorand functionality. The second research question (RQ.2) explores a suitable method for implementing a new language inStorm. It is suggested to first implement a foundational structure comprising of statements, blocks,scope handling and variable declarations. From this foundation, new functionalities can be graduallyintroduced and tested by connecting them to the appropriate location in the structure. When allfunctionality is added and tested a refactoring step can take place to modify the BNF if needed.