DSpace Coleção:http://www.repositorio.ufc.br/handle/riufc/3842023-03-22T04:01:09Z2023-03-22T04:01:09ZGeração de malhas de alta ordem para análise isogeométrica utilizando elementos de Bézier racionaisBarroso, Elias Saraivahttp://www.repositorio.ufc.br/handle/riufc/688782022-10-18T17:14:34Z2022-01-01T00:00:00ZTítulo: Geração de malhas de alta ordem para análise isogeométrica utilizando elementos de Bézier racionais
Autor(es): Barroso, Elias Saraiva
Abstract: Isogeometric Analysis (IGA) is a numerical method that has been receiving increasing attention in the last decade. The main goal of IGA is to closely couple geometric modeling with numerical analysis in Computer-Aided Engineering (CAE) systems. To that end, in IGA both components use the same geometric representation, e.g., Bézier curves and surfaces, NURBS, and T-Splines. However, in many cases, especially in CAE systems based on the Finite Element Method, the geometric representation in a CAD system cannot be directly employed in numerical simulation, as only the boundary of the geometry is parametrized. In these cases, an interior parametrization must be constructed before performing isogeometric analysis. The domain parametrization techniques for NURBS and T-Splines are not enough robust to handle complex geometries, e.g. geometries with multiple holes and narrow regions. In addition, the use of trimmed surfaces in IGA has considerable complexity, either in the construction of their parameterization or when these models are employed directly in numerical simulations. On the other hand, the use of high-order unstructured mesh generator algorithms to construct domain parametrization is more robust in these cases. This work presents an algorithm for generation of unstructured geometrically exact meshes composed of rational Bézier triangles of arbitrary degree and applies it to plane models described by NURBS curves in a Boundary-Representation (B-Rep) scheme. The proposed algorithm is capable of generating high quality coarse meshes even when high curvature segments are considered. The proposed algorithm attains superior performance when compared to a well-known algorithm in the literature and produces meshes with similar quality in comparison to
meshes obtained through quality optimization. The algorithm is generalized to surface models in 3D, where geometry is given by trimmed NURBS. Structured meshing algorithms for generation of rational Bézier triangles and quadrilaterals are also presented. Moreover, a shell formulation based on degenerated solid for rational Bézier elements is presented. The algorithms developed are used in elasticity and heat transfer problems, and static and free vibration analysis of shells, using the elements developed in this work. The performance of rational Bézier elements is assessed in several numerical examples, demonstrating its convergence under mesh refinement.2022-01-01T00:00:00ZResultados em jogos de coloração e perseguição em grafosCosta, Eurinardo Rodrigueshttp://www.repositorio.ufc.br/handle/riufc/684642022-09-22T15:34:44Z2022-01-01T00:00:00ZTítulo: Resultados em jogos de coloração e perseguição em grafos
Autor(es): Costa, Eurinardo Rodrigues
Abstract: In this thesis we study three games in graphs. In the graph coloring game we answer a long-standing open question proposed by Bodlaender in 1991: the game chromatic number is PSPACE-hard. In the greedy coloring game we also prove that the game Grundy number is PSPACE-hard. In fact, we prove that both problems, the graph coloring game and the greedy coloring game, are PSPACE-Complete even if the number of colors is the chromatic number. Moreover, we prove that the game Grundy number is equal to the chromatic number for several superclasses of cographs, extending a result of Havet and Zhu in 2013. In the spy game we obtain an upper bound on the strong product of two general graphs and obtain examples of King grids that match this bound and other examples for which the guard number is smaller. We also obtain the exact value of the guard number in the lexicographic product of two general graphs for any distance d ≥ 2. From the algorithmic point of view, we prove a positive result: if the number k of guards is fixed, the spy game is solvable in polynomial time O(n3k+2) for every speed s ≥ 2 and distance d ≥ 0. In other words, the spy game is XP when parameterized by the number of guards. This XP algorithm is used to obtain an FPT algorithm on graphs with few P4’s. As a negative result, we prove that the spy game is W[2]-hard even in bipartite graphs when parameterized by the number of guards, for every speed s ≥ 2 and distance d ≥ 0, extending the hardness result of Cohen et al. in 2018.2022-01-01T00:00:00ZLocation prediction from external sensors trajectoriesCruz, Lívia Almadahttp://www.repositorio.ufc.br/handle/riufc/681742022-09-13T15:22:32Z2022-01-01T00:00:00ZTítulo: Location prediction from external sensors trajectories
Autor(es): Cruz, Lívia Almada
Abstract: This thesis proposes a multi-task deep learning-based scheme to predict the next location from trajectories captured by external sensors (eg traffic surveillance cameras or speed cameras). The positions reported in these trajectories are sparse, due to the distribution of the sensors, and incomplete because the sensors can fail to register the passage of objects. This framework includes different pre-processing steps to align the representation of trajectories and deal with the problem of missing data. We present a multitasking learning approach based on recurrent neural networks. This approach uses time and space information in the training phase to learn more meaningful representations. The multi-task learning model jointly with the pre-processing step substantially improves the prediction performance. This thesis also deals with the problem of representation learning for trajectory data. Representation learning concerns the problem of learning low-dimensional representation from complex data, and it is an essential task in machine learning. We evaluate how natural language processing models capture the representation of sensors and trajectories. The empirical evaluation shows that the space of features identified by such models can capture the spatial similarity relationships for sensors and trajectories within a given neighborhood. We also evaluate how these representations improve a location prediction model.2022-01-01T00:00:00ZMain memory database instant recoveryAraújo, Arlino Henrique Magalhães dehttp://www.repositorio.ufc.br/handle/riufc/669022022-07-04T19:17:47Z2022-01-01T00:00:00ZTítulo: Main memory database instant recovery
Autor(es): Araújo, Arlino Henrique Magalhães de
Abstract: Main Memory Databases (MMDBs) technology handles the primary database in Random Access Memory (RAM) to provide high throughput and low latency. However, volatile memory makes MMDBs much more sensitive to system failures. The contents of the database are lost in these failures, and, as a result, systems may be unavailable for a long time until database recovery process has been finished. Therefore, novel recovery techniques are needed to repair crashed MMDBs as quickly as possible. This thesis presents MM-DIRECT (Main Memory Database Instant RECovery with Tuple consistent checkpoint), an MMDBs recovery technique able to schedule transactions simultaneously with the database recovery process at system startup. Thus it is giving the impression that the database is instantly restored. The approach implements a tuple-level consistent checkpoint to reduce the recovery time. In order to validate the proposed approach, experiments have been performed in a prototype implemented on the Redis database. The results show that the instant recovery technique effectively provides high transaction throughput rates even during both the recovery process and normal database processing.2022-01-01T00:00:00Z