DSpace Coleção:

Alternativas para um grande consumidor no mercado de energia elétrica no Brasil: análise de cenários e viabilidade sob incertezas utilizando modelagem estocástica

Mon, 01 Jan 2024 00:00:00 GMT

Título: Alternativas para um grande consumidor no mercado de energia elétrica no Brasil: análise de cenários e viabilidade sob incertezas utilizando modelagem estocástica Autor(es): Melo, Bruno Valdivino Abstract: Many countries worldwide have made efforts to reduce greenhouse gas emissions and develop environmental, social, and governance (ESG) practices that can guide sustainability-focused investments. Diversifying a country's energy matrix through the incorporation of alternative renewable sources is strongly associated with sustainable development and opportunities for improvements in ESG metrics. Given its abundant natural resources, it is expected that the industry in Brazil will continue to benefit from the growth of the renewable energy sector. Concurrently, the evolution of the Energy Market in Brazil has provided financial possibilities that directly impact key sectors driving the economy. One of the options is the alternative of self-electricity production, which may offer benefits to the consumer. However, due to the complexities of the variables inherent in the electricity sector, it is necessary to analyze existing financial risks. This study aims to present the alternatives offered by the Brazilian electricity market to large electricity consumers and analyze the feasibility and risks for consumer classification in possible scenarios within the Free Contracting Environment (ACL): long-term power purchase agreement (PPA) and possible arrangements within the context of energy self-production, such as self-production through own investment, self-production through equivalence, and self-production through leasing. For the analyzed scenarios, uncertainties, risks, and fluctuations inherent to the factors comprising the characteristics of the Energy Market in Brazil are considered. To achieve this, stochastic characteristics are applied to the variables that most impact the models, aiming to optimize and provide greater support to the consumer for decision-making. Sensitivity analyses and feasibility metrics are applied for comparative purposes between scenarios. For the analyzed situation, it is observed that the most probable Net Present Value (NPV) for the self-production through own investment alternative can result in savings of approximately 32.9% compared to the PPA contract. The most likely discounted payback found was 10.32 years. Furthermore, the observed Internal Rate of Return (IRR) for all scenarios demonstrated feasibility for investment in self- generation. Optimizations are conducted, returning essential values in the decision-making process regarding the contracting of self-production models through equivalence or leasing. While self-generation investment has proven advantageous, given the adopted premises, it is crucial for the investor to assess the impacts of capital expenditure (CAPEX) within the business context, as it represents a potentially relevant value. Tipo: Dissertação

Diagnóstico multimodal da doença de parkinson com arquitetura agentica colaborativa baseada em internet de modelos de linguagem médica

Thu, 01 Jan 2026 00:00:00 GMT

Título: Diagnóstico multimodal da doença de parkinson com arquitetura agentica colaborativa baseada em internet de modelos de linguagem médica Autor(es): Peixoto Junior, Eugenio Abstract: This thesis proposes an intelligent platform to support the diagnosis of Parkinson’s Disease (PD), integrating advanced technologies such as the Internet of Medical Things (IoMT), Multimodal Retrieval-Augmented Generation (RAG), efficient model compression, and medical language model operations (LLMOps and VLMOps). Unlike previous approaches that rely on isolated models or single data modalities, this work defines a clinical agentive ecosystem composed of autonomous and collaborative agents powered by Large or Small Language Models (LLMs/SLMs), with reinforcement learning (RL). The system architecture is designed to be adaptable, explainable, and patient-centered, with an emphasis on reliability, fast response times, and practical deployment in real clinical environments. Heterogeneous data modalities, including clinical text, medical imaging, sensor signals, and voice recordings, are fused through advanced RAG variants such as Multimodal RAG, GraphRAG, and Agentic RAG, enabling contextualized inference without the need for extensive fine-tuning. The methodology includes evaluating computational cost and operational efficiency in simulated production environments, using metrics such as GPU utilization, LLM call cost, and infrastructure scalability. Compression techniques such as 4-bit quantization and QLoRA are employed to enable robust models to run on edge devices. Finally, intelligent agents refine their inference strategies through post-training with reinforcement learning, using rewards derived from clinical analyses to improve both the accuracy and practical applicability of the system. Experimental results showed competitive performance across classifiers, with Gradient Boosting achieving 83.85% accuracy and an F1-score of 88.74%, and HistGradient Boosting reaching 82.92% accuracy, an F1-score of 88.12%, and an AUC-ROC of 0.909. While models such as SVM and KNN demonstrated higher sensitivity in some cases, they also showed a greater rate of false positives, highlighting the superior balance of the proposed approach. This work stands out for its capacity to integrate with real clinical workflows and holds strong potential to serve as a reference model for neurointelligent platforms in connected digital healthcare. Tipo: Tese

Arquitetura neuro-simbólica com floresta de caminhos ótimos para classificação de EEG através dos macroestados com GRPO

Thu, 01 Jan 2026 00:00:00 GMT

Título: Arquitetura neuro-simbólica com floresta de caminhos ótimos para classificação de EEG através dos macroestados com GRPO Autor(es): Guimarães, Raniere Rocha Abstract: Neurodegenerative dementias represent a growing challenge for early diagnosis, requiring computational methods capable of capturing dynamic changes in the functional organization of brain networks. This work proposes a neuro-symbolic artificial intelligence architecture that combines unsupervised Optimum-Path Forest (OPF) with Deep Neural Network (DNN) and a decision strategy grounded in Group Relative Policy Optimization (GRPO), applied to the classification of electroencephalography (EEG) signals through microstates. OPF is employed to symbolically model the structure of microstate temporal parameters, while the DNN captures non-linear relationships and complex discriminative patterns. The combination of predictions is performed using a decision mechanism inspired by GRPO, which dynamically adjusts the relative contributions of each model based on their performance. The study used a dataset from AHEPA University General Hospital, consisting of resting-state EEG records from Cognitively Normal (CN), patients with Alzheimer’s Disease (AD), and patients with Frontotemporal Dementia (FTD). The methodology includes preprocessing stages, microstate extraction through modified k-means, a systematic evaluation of 37 distance measures in the context of OPF, and stratified cross- validation. Results demonstrate that the proposed architecture achieves an average accuracy of 97.31%±0.58 and an F1-score of 97.34%±0.58, even under class imbalance conditions, outperforming traditional approaches such as k-means, Self-Organizing Map (SOM), and isolated neural classifiers. The use of a relative decision policy improves performance stability and consistency in distance-measure rankings. The main contributions of this work include: (i) a neuro-symbolic architecture for EEG microstate classification; (ii) an extensive comparative analysis of distance measures in the OPF context; and (iii) the application of a GRPO-based decision strategy for adaptive model fusion. The results indicate relevant translational potential for clinical applications supporting the differential diagnosis of neurodegenerative dementias. Tipo: Tese

Microwave brain stimulation plarform: neural network-based parameter optimization for non-invasive applications

Wed, 01 Jan 2025 00:00:00 GMT

Título: Microwave brain stimulation plarform: neural network-based parameter optimization for non-invasive applications Autor(es): Pereira, Francisco Estevão Simão Abstract: Non-invasive brain stimulation has numerous fundamental challenges in balancing penetration depth and spatial focus. Established methods, such as Transcranial Magnetic Stimulation (TMS) and Transcranial Direct Current Stimulation (tDCS), are effective only on superficial cortical targets. Microwave Brain Stimulation (MBS) emerges as a promising alternative to reach subcortical targets, but its clinical viability critically depends on the complex and multiobjective optimization of the antenna arrangement. This modeling problem, which involves the simultaneous adjustment of physical (hardware) and electrical (beam control) parameters, remains a gap in the literature. This thesis proposes and validates a new computational platform, the Hybrid Coevolutionary Method (HCM-TBS), specifically designed to solve this problem. The platform implements a multi-stage modular architecture that integrates global search algorithms (PSO), diversity and robustness (GA, SA), and a coevolutionary optimization core.This core decomposes the problem into subpopulations of physical (hardware) and electrical (software) parameters, optimizing them cooperatively through joint fitness assessments in an FDTD (Meep) simulator. The performance of the HCM-TBS platform was rigorously validated against baseline methods thru 30 independent runs for each group. The results demonstrate that HCM-TBS achieves statistically superior beamforming accuracy (Directivity) (p < 0.001). Moreover, the analysis reveals that the proposed platform produces solutions with significantly lower variability, demonstrating high robustness and reliability, in contrast to the extreme performance instability observed in the baseline methods. In light of the findings, this thesis contributes a robust and statistically validated computational framework for the modeling of non-invasive neurostimulation systems, prioritizing reliability and precision of focus, essential steps for future clinical translation. Tipo: Tese