http://repositorio.ufc.br/handle/riufc/22000 2026-04-05T13:59:10Z http://repositorio.ufc.br/handle/riufc/83081 Título: Algoritmos de ramificação e poda aprimorados por aprendizado profundo na resolução do problema da clique máxima Autor(es): Sousa, Igor Alan Albuquerque de Abstract: This work addresses the Maximum Clique Problem (MCP) in graphs, a well-known challenge in graph theory and combinatorial optimization. Exact algorithms are computationally expensive, and heuristic algorithms designed for MCP rely on good initial solutions. Therefore, the application of machine learning (ML) methods to solve the MCP is considered. The main goal is to use ML models to decide the branching order for Branch and Bound (B&B) algorithms that solve the MCP. Initially, the MCP and B&B algorithms are addressed. Subsequently, neural networks are presented as a powerful tool for learning branching patterns in search trees. Finally, the results of integrating ML with B&B algorithms that solve the MCP are presented. Tipo: TCC 2025-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/83080 Título: Análise comparativa de gerenciadores de pacotes JavaScript: NPM, PNPM e Yarn Autor(es): Lima, Rayrisson Vinicius Alves de Abstract: With the increasing complexity in the development of modern JavaScript applications, the use of package managers has become essential to handle the installation, updating, and organization of dependencies. In this context, tools such as NPM, Yarn, and PNPM have emerged as widely adopted alternatives, each offering distinct approaches that directly impact developers’ workflows. This work aims to perform a comparative analysis of the NPM, PNPM, and Yarn package managers, evaluating their characteristics, strategies, and behavior regarding aspects such as performance, storage usage, security, and compatibility. For this purpose, a practical project with a high number of dependencies was developed, and a series of experiments were conducted on different operating systems, focusing on objective metrics and empirical observations. The proposed analysis seeks to provide a practical guide to help developers choose the tool that best suits the specific needs of their projects. Tipo: TCC 2025-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/83079 Título: Verificação de autenticidade de contas de órgãos públicos em mídias sociais com blockchain: uma abordagem para transparência e confiabilidade Autor(es): Coutinho, Roberto de Oliveira Abstract: With the growing presence of public agencies on social media, the risks related to the dissemination of false information by unofficial profiles increase. This work proposes the development of a web system that uses blockchain technology to guarantee the authenticity of institutional profiles, linking them in an auditable manner to their respective public agencies. The application was built with the Hyperledger Fabric platform and offers features such as registering institutions, linking and revoking profiles, and public consultation to verify these links. To assess the system’s acceptance, a user study was conducted, based on the TAM (Technology Acceptance Model). The results indicated that 86.6% of participants fully or partially agree that the system is useful to avoid deception caused by fake profiles and would use it regularly. Most users demonstrated ease in using the application, even with little prior knowledge of blockchain. Thus, the solution proved to be effective in promoting greater trust and transparency in digital communication between public agencies and society. Tipo: TCC 2025-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/83078 Título: Avaliação de plataformas de contêineres AWS para offloading computacional de aprendizado de máquina em dispositivos de baixo poder computacional Autor(es): Rosendo Filho, Eliton Lima Abstract: The growing demand for Machine Learning (ML) applications on devices with limited computing power poses significant performance and energy consumption challenges. This work addresses this issue through computational offloading, evaluating the feasibility and efficiency of cloud container platforms for executing ML inference tasks. The main objective is to compare two Amazon Web Services (AWS) approaches: the serverless AWS ECS Fargate service and running containers on AWS ECS managed instances with EC2. For the analysis, a prototype image classification application is developed using the MobileNetV3Large model, and systematic experiments are conducted to collect metrics on latency, total processing time, and financial cost. The results demonstrate that, although the EC2 approach presents marginally superior performance in terms of latency, Fargate stands out as the more cost-effective option with lower operational complexity. It is concluded that Fargate represents a robust and financially viable solution for ML offloading scenarios, offering an advantageous trade-off between cost, simplicity, and adequate performance for most applications. Tipo: TCC 2025-01-01T00:00:00Z