http://repositorio.ufc.br/handle/riufc/384 2026-11-15T08:37:30Z 2026-11-15T08:37:30Z Nunes Júnior, Francisco Carlos Freire http://repositorio.ufc.br/handle/riufc/88213 2026-10-06T16:33:54Z 2026-01-01T00:00:00Z

Título: From hot spot street segments to deployable patrol routes: an approach for creating crime-weighted routing on urban networks Autor(es): Nunes Júnior, Francisco Carlos Freire Abstract: Hot spot policing depends on a visible preventive presence, yet effectively converting street segments with a high historical crime record into foot patrol routes remains an open operational challenge. We propose Hot-segment-Aware Route-set Planning (HARP), an end-to-end approach that assigns georeferenced incidents to street segments through a crime-aware map-matching procedure, identifies hot segments, generates feasible circular patrol routes within distance-bounded street-network neighborhoods, and composes a plan of up to k edge-disjoint routes under minimum and maximum walking-distance constraints. HARP separates route construction from route-set selection, so that both stages can be exchanged, compared, and combined. Four interchangeable constructors are provided: HotSee, a fast heuristic that links hot segments through shortest-path connectors; HotStar, an A∗ -guided heuristic driven by accumulated crime weight; BEPaR, a budget-constrained exhaustive enumerator; and ARSA, a classifier-guided constructor that predicts whether exhaustive enumeration remains affordable in each local search space. The final route set is obtained either sequentially, with candidate regeneration after each selection, or through an Integer Linear Programming (ILP) formulation of the problem as weighted set packing, optionally preceded by a candidate-enrichment stage. Experiments on five real-world study areas in two countries show that joint optimization alone is not sufficient: the fixed-pool ILP was outperformed by sequential selection in most combinations, whereas the enriched ILP obtained the highest crime coverage in 14 of 15 combinations. Under this procedure no single constructor dominated, since BEPaR led in three areas and HotStar in the other two, remaining within 1.60 percentage points at one to two orders of magnitude less runtime, while ARSA matched or exceeded BEPaR coverage and reduced runtime by factors of 2 to 10 where enumeration was costly. Combining candidate pools increased coverage in every area, by up to 1.97 percentage points over BEPaR. These results indicate that patrol-plan quality depends on the interaction between candidate construction and route-set selection, and that HARP supports configurations adaptable to different operational priorities and computational budgets. Tipo: Tese

2026-01-01T00:00:00Z Aragão Júnior, Belmondo Rodrigues http://repositorio.ufc.br/handle/riufc/88212 2026-10-06T13:22:55Z 2025-01-01T00:00:00Z

Título: Milestone: framework para monitoramento contínuo de desempenho em sistemas auto-adaptativos. Autor(es): Aragão Júnior, Belmondo Rodrigues Abstract: Self-adaptive systems autonomously modify their behaviour in response to changes in the environment, using an iterative process called MAPE-K (monitoring, analysis, planning, execution and knowledge). One of the biggest challenges in developing these systems is monitoring and evaluating non-functional requirements, which are fundamental to their adaptive capabilities. To support adaptation decisions, these requirements must be treated as entities at runtime, capable of being collected, analysed and synthesised throughout the self-adaptation process. In this context, Milestone is proposed, which is a framework focused on performance adaptability, designed based on the MAPE-K management cycle and using neural networks as a central computational model. An experimental implementation is done to validate the proposed framework, specifically in the analysis, planning and knowledge stages. The results demonstrated an analysis accuracy of 85%, a ROC-AUC value of 0.86 and a loss function of 0.43, indicating high accuracy in detecting performance anomalies. The planning phase achieved a latency of less than 1 second and a success rate of 66.67%, while the knowledge phase achieved 100% effective utilisation. Furthermore, the results indicate adaptability in variable environmental conditions, with high planning accuracy and minimum latency, demonstrating the effectiveness of the framework in managing the performance non-functional requirement. Tipo: Tese

2025-01-01T00:00:00Z Silva, Bruno Correia da http://repositorio.ufc.br/handle/riufc/87982 2026-09-18T11:51:00Z 2026-01-01T00:00:00Z

Título: Procedural content generation approach with generative artificial intelligence for location based games Autor(es): Silva, Bruno Correia da Abstract: Location-based games (LBG) are characterized by gameplay mechanics directly influenced by players’ physical locations. Early LBG relied on manually constructed environments and live actors to create immersive settings, while players interacted via pre-scripted quests and devices. In contrast, contemporary designs have a decreasing reliance on real-world contextual integration, shifting personalization from real-world contexts to virtual spaces. The automation of content generation in game development has advanced alongside AI innovations, exemplified by roguelike games, which leverage procedural algorithms for dynamic gameplay. AI-driven technologies, particularly large language models (LLM), have become integral to the gaming industry. Trained on vast datasets, LLM are capable of generating text, simulating dialogues, and producing adaptive narratives, which are critical functions for enhancing player engagement. Some games in the industry use LLM as algorithms for procedural content generation (PCG). One example is narrative generation and level rules to facilitate replayability. Within the gaming industry’s framework, LBG often underutilize real-world data, employing physical locations merely as gameplay channels rather than integrating contextual information. These systems eliminate the need for live actors or fixed scripts but lack content generation tailored to specific locales, such as location-adaptive narratives. This transition from hybrid to virtual immersion raises questions about trade-offs in design complexity and player experience. Leveraging LLM’ generative processing capabilities and accessibility offers opportunities to merge real-world elements into gameplay. This approach enables personalized content integration, enhancing user interaction with their context. This study investigates the potential of generative AI (GenAI), particularly LLM, integrates design LBG through such as PCG. Motivated by gaps in PCG rese arch within LBG, a literature review employing identified key research questions and conceptual frameworks was conducted. Challenges such as reduced designer involvement and scalable geographic personalization were identified. The primary aim was to design, implement, and evaluate a PCG system for LBG that integrates GenAI to generate context-aware content. This study introduces Top Triumphs, a collectible card game where card attributes are dynamically generated based on location-specific variables. This work contributes by integrating GenAI into pervasive gaming (LBG), demonstrating how LLM can expand content reuse while incorporating real-world details with minimal human oversight. By addressing these challenges, the research advances the intersection of AI-driven design and immersive gameplay where it brings enhancing virtual immersion through real-world integration. Tipo: Tese

2026-01-01T00:00:00Z Gomes, Francisco Anderson de Almada http://repositorio.ufc.br/handle/riufc/86582 2026-08-27T12:05:39Z 2026-01-01T00:00:00Z

Título: Uma abordagem estruturada para observabilidade em microsserviços: taxonomia, catálogo e framework de detecção de anti-padrões Autor(es): Gomes, Francisco Anderson de Almada Abstract: Software systems increasingly rely on microservices-based architectures to enhance scalability, modularity, and continuous deployment. Although this approach simplifies development by promoting a functional decomposition of components, it also introduces significant operational complexity, making failures more frequent and harder to diagnose. In this context, observability emerges as a fundamental concept, defined as the ability to understand and diagnose the internal behavior of a system based on its external outputs, such as metrics, logs, and traces. However, despite its importance, observability is often poorly implemented due to the lack of standardized practices, resulting in ineffective monitoring, alert fatigue, and low efficiency in incident response. Although existing studies discuss concepts, tools, and challenges related to observability, no prior work has focused specifically on observability anti-patterns, recurrent practices that undermine monitoring effectiveness, nor proposed solutions capable of detecting them automatically. Furthermore, there is a lack of a comprehensive taxonomy to classify and organize existing studies on observability. To address these gaps, this thesis presents a taxonomy focused on observability in microservices-based applications, constructed through a systematic mapping of the literature. A total of 84 relevant studies published between 2019 and 2025 were analyzed, providing a comprehensive overview of the field. The review also identifies tools, benchmarking applications, and real-world datasets used in the selected studies. Complementing this contribution, the thesis develops a systematized catalog of observability anti-patterns, offering an approach to identify and mitigate harmful practices. This catalog serves as a practical guide to support teams in building more reliable and efficient systems. In total, 37 anti-patterns were identified, whose relevance was evaluated by 60 experts, achieving an agreement rate of 95%. Finally, the thesis introduces the Observa framework, designed to automatically detect observability anti-patterns. Its operation was evaluated through three experiments and a proof of concept, which demonstrated its technical feasibility and approved its adoption, receiving a recommendation of excellence from the evaluators. Tipo: Tese

2026-01-01T00:00:00Z