http://repositorio.ufc.br/handle/riufc/23978 2026-05-15T07:49:28Z http://repositorio.ufc.br/handle/riufc/86038 Título: Métodos de aprendizado de máquina aplicados ao monitoramento ambiental Autor(es): Aragão Neto, Pedro Sousa de Abstract: In recent years, the world has been undergoing enormous environmental transformations and where there is a constant occurrence of major technological innovations, the need arises to understand how technology can be beneficial to the environment. Neural Networks have emerged as powerful tools for classifying complex data, offering solutions for identifying patterns and anomalies. This article explores the application of three neural networks – Artificial Neural Networks (ANN), Multilayer Perceptron (MLP) and Radial Basis Function Networks (RBF) – in the context of urban environmental data classification. We present a comparative analysis of their performance using real datasets, highlighting their strengths and limitations in accurately categorizing environmental parameters. Our results demonstrate the effectiveness of these neural network models in contributing to environmental monitoring, providing valuable insights. Tipo: TCC 2026-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/86008 Título: Análise comparativa de ferramentas de CI/CD: um estudo de caso da migração entre GitHub Actions e GitLab CI no sistema Adote Fácil Autor(es): Gomes, Márcio Bruno Loiola Abstract: The adoption of DevOps practices and the implementation of Continuous Integration and Continuous Delivery (CI/CD) pipelines have become fundamental pillars for ensuring agility and reliability in modern software development. However, the diversity of tools available on the market imposes significant interoperability challenges. Choosing between leading solutions, such as GitHub Actions and GitLab CI/CD, involves technical trade-offs that directly impact team productivity. This work aims to conduct a practical comparative analysis between these tools, documenting the pipeline migration process of the “Adote Fácil” application, a microservices-based pet adoption platform. To this end, a complete portability of the automation flow was performed, identifying critical challenges such as syntactic differences (workflow-centric versus stage-centric approach) and the complexity of Docker-in-Docker (DinD) orchestration. Quantitative results demonstrated that migrating to GitLab CI resulted in a 2.6-fold increase in configuration code verbosity and required 7 correction iterations for complete stabilization. Qualitatively, it is evidenced that while GitHub Actions prioritizes ease of use through abstractions, GitLab CI offers greater granular control over the execution environment, albeit requiring a steeper learning curve. The study contributes a technical guide for developers facing transition challenges between automation tools. Tipo: TCC 2026-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/86007 Título: BuriedBrains: um ambiente multiagente inspirado em roguelike para avaliação de memória e tomada de decisão Autor(es): Silva, Ismael Soares da Abstract: This work presents the development and validation of BuriedBrains, a multi-agent simulation environment based on procedural generation and partial observability, structured through hybrid dynamic graphs. The primary objective is to measure the performance of Reinforcement Learning (RL) agents in solving long-term planning and generalization problems under conditions of permadeath and resource constraints. The environment’s topology, inspired by roguelike mechanics — specifically the title Buriedbornes (Nussygame, 2016) —, utilizes Directed Acyclic Graphs (DAGs) for Progression Zones and cyclic graphs for Sanctuaries. These structures demand inventory and skill management, as state transitions result in persistent changes to variables such as Karma and the agent’s survival probability. The methodology consisted of a comparative study to isolate the impact of temporal recurrence, contrasting a reactive architecture (PPO) with a recurrent one (LSTM). Both models were refined through Hyperparameter Optimization (HPO) and integrated with Self-Attention mechanisms. Data suggests that the characteristics of BuriedBrains restrict the performance of reactive models under the tested conditions, as reflected by PPO’s stabilization at a 65% Explained Variance (EV) plateau, which remained constant despite increases in network complexity or sampling horizon. The recurrent agent demonstrated 24% superior performance in HPO tests and reached 80% explained variance. Results suggest the architecture was capable of sustaining policies focused on attribute specialization and exhibiting multi-agent interaction patterns. Visualization-based analysis identified convergence toward reciprocal non-aggression states, where memory enabled the optimization of episode longevity over zero-sum interactions. BuriedBrains thus serves as a benchmark for evaluating long-term temporal dependency resolution and decision-making in autonomous systems. Tipo: TCC 2025-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/86004 Título: Integração e aprimoramento de aplicativos para acolhimento, acessibilidade e suporte acadêmico no ambiente universitário Autor(es): Sant’anna, Ionara Brandão Abstract: Access to institutional and academic information can represent a significant challenge for university students, especially those entering higher education for the first time. Difficulties in obtaining information about scholarships, course structures and accessibility tend to compromise their adaptation and permanence in the university. In this context, the present work has as objective the integration and the improvement of two applications developed within the scope of the UFC: Um Estranho no Ninho and Mapa Acessível. The first gathers relevant academic information, especially for students newly entering higher education. The second has as its focus the mapping of the accessibility points of the campus, in addition to informational content. Considering the broad familiarity of students with Um Estranho no Ninho and the range of information it provides, the application was chosen as the functional basis of the new version. From this structure, Mapa Acessível was incorporated as one of its main functionalities, considering its relevance in providing essential information about accessibility on campus. However, in view of the technical limitations of the no-code environment of the original application, it is opted for the development of the system from scratch, in an open-code environment, using modern technologies such as React and TypeScript, and structuring the solution as a Progressive Web App (PWA), which combines benefits of web and native applications. The methodology focuses on three axes: development of the new version in an architecture prepared to support the incorporation of other campuses, integration between the systems, and the implementation of accessibility resources, in accordance with the guidelines of the Web Content Accessibility Guidelines (WCAG), which stipulate digital accessibility standards. As a result, a responsive and inclusive platform is obtained, which contributes to the strengthening of student permanence and the democratization of access to information. Tipo: TCC 2026-01-01T00:00:00Z