Ciencias Exactas y Ciencias de la Salud
Permanent URI for this collectionhttps://hdl.handle.net/11285/551014
Pertenecen a esta colección Tesis y Trabajos de grado de los Doctorados correspondientes a las Escuelas de Ingeniería y Ciencias así como a Medicina y Ciencias de la Salud.
Browse
Search Results
- A generalist reinforcement learning agent for compressing multiple convolutional neural networks(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12-11) González Sahagún, Gabriel; Conant ablos, Santiago Enrique; emipsanchez; Ortíz Bayliss, José Carlos; Cruz Duarte, Jorge Mario; Gutiérrez Rodríguez, Andrés Eduardo; School of Engineering and Sciences; Campus MonterreyDeep Learning has achieved state-of-the-art accuracy in multiple fields. A common practice in computer vision is to reuse a pre-trained model for a completely different dataset of the same type of task, a process known as transfer learning, which reduces training time by reusing the filters of the convolutional layers. However, while transfer learning can reduce training time, the model might overestimate the number of parameters needed for the new dataset. As models now achieve near-human performance or better, there is a growing need to reduce their size to facilitate deployment on devices with limited computational resources. Various compression techniques have been proposed to address this issue, but their effectiveness varies depending on hyperparameters. To navigate these options, researchers have worked on automating model compression. Some have proposed using reinforcement learning to teach a deep learning model how to compress another deep learning model. This study compares multiple approaches for automating the compression of convolutional neural networks and proposes a method for training a reinforcement learning agent that works across multiple datasets without the need for transfer learning. The agents were tested using leaveone- out cross-validation, learning to compress a set of LeNet-5 models and testing on another LeNet-5 model with different parameters. The metrics used to evaluate these solutions were accuracy loss and the number of parameters of the compressed model. The agents suggested compression schemes that were on or near the Pareto front for these metrics. Furthermore, the models were compressed by more than 80% with minimal accuracy loss in most cases. The significance of these results is that by escalating this methodology for larger models and datasets, an AI assistant for model compression similar to ChatGPT can be developed, potentially revolutionizing model compression practices and enabling advanced deployments in resource-constrained environments.
- The impact of loading-unloading zones for freight vehicles on the last-mile logistics for nanostores in emerging markets(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12-11) Mora Quiñones, Camilo Andrés; Cárdenas Barrón, Leopoldo Eduardo; emimmayorquin; Fransoo, Jan C.; Smith Cornejo, Neale Ricardo; Loera Hernández, Imelda de Jesús; School of Engineering and Sciences; Campus Monterrey; Veláaquez Martínez, Josue CuauhtémocEvery year, more than 26 billion deliveries are made globally to serve nanostores, the largest grocery retail channel in the world. At each stop, company representatives face a persistent challenge: finding a place to park. While the problem seems simple, it is remarkably complex and far from easy to solve. In emerging markets, where cities have grown rapidly and often without proper planning, fragmented markets and inadequate infrastructure exacerbate the issue. Multiple stakeholders compete for limited curb space, and the lack of dedicated parking disrupts last-mile efficiency, forcing drivers to either cruise for parking or resort to illegal parking. These behaviors lead to increased vehicle emissions, noise pollution, and additional costs. This dissertation provides key insights into last-mile logistics for nanostores in emerging markets, contributing to academic literature and offering practical implications to address the parking problem. The first study addresses the parking challenges faced by freight vehicles serving nanostores, identifying key factors affecting dwell time efficiency and suggesting operational improvements. In the next study, the focus shifts to the implementation of Loading-Unloading Zones (LUZs) as a targeted intervention, analyzing their impact on reducing air and noise pollution in urban areas. The last study extends this analysis by exploring the effects of LUZs on traffic flow, evidencing how their introduction can improve vehicle speed and reduce congestion in densely populated city streets. Together, these studies provide a detailed exploration of the operational, environmental, and infrastructural challenges of last-mile logistics, while offering concrete strategies to improve urban logistics in emerging markets. This dissertation contributes by expanding the body of knowledge and offering actionable managerial insights with the potential to drive meaningful impact. These include enhancing air quality, reducing noise pollution, lowering carbon emissions, improving traffic flow, and achieving substantial cost savings for companies distributing goods to nanostores in emerging markets.
- Environmental monitoring to estimate indoor occupancy levels based on Semi-supervised machine learning and data fusion for building management(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12-03) Mena Martínez, Alma Rosa; Ceballos Cancino, Hector Gibran; emipsanchez; Alvarado Uribe, Joanna; Cantu, Francisco J.; García, Juan Pablo; School of Engineering and Sciences; Campus Monterrey; Schmitt, JanOccupancy information is essential for space management, energy efficiency, and in times of the COVID-19 pandemic, for crowd control. Obtaining labeled data is challenging due to hardware limitations, privacy considerations, and the required underlying costs. Furthermore, venues over 200 m2 require data fusion techniques. Therefore, this thesis mainly focuses on exploring the potential of Semi-Supervised Learning (SSL), which only needs a few labeled data and a large amount of unlabeled data, to estimate the occupancy levels in enclosed spaces. This study presents an empirical comparison between Supervised ML and SSL models as well as data fusion techniques in real-life university classrooms and offices (uncontrolled conditions) at the University of the West of England, Bristol, UK, and Tecnologico de Monterrey, Mexico. The data was collected for three weeks at each scenario using an in-house developed Internet of Things (IoT) device that measures air temperature, relative humidity, and atmospheric pressure. The ground truth records were gathered through manual logging of occupancy levels. Datasets’ sizes averaged 2350 entries with only 280 labeled instances per dataset. Support Vector Machine (SVM), Random Forest (RF), and Multi-Layer Perceptron (MLP) were used to define a performance baseline for supervised ML. Self-Training (ST) and Label Propagation (LP) were tested for SSL. In addition, several feature fusion methods were explored, including Chi-squared, ANOVA F-test, Spearman and Kendall’s Tau correlation, Mutual Information, Averages, Recursive Feature Elimination, and Principal Component Analysis. The models were evaluated using Accuracy, Precision, Recall, F1-score, Confusion Matrix, and High - Quality Supervised Baseline. ST achieved superior performance compared to baseline models (SVM, RF, MLP) with a highest average accuracy of 90.96% compared to SVM (86.66%). Furthermore, the data fusion results indicated that the Chi-squared approach for feature fusion outperformed others with an F1-score average of 95% and an accuracy average of 99%. These results demonstrate the effectiveness of SSL for indirect occupancy estimation while reducing the need for extensive data collection and labeling.
- Optimization and sustained release of green lentil polyphenols through instant controlled pressure drop and encapsulation in PLGA nanoparticles(2024-12-03) Tienda Vázquez, Mario Adrián; Almanza Arjona, Yara C.; emimmayorquin; Cardador Martínez, Anabertha; Quintus Scheckhuber, Christian; Téllez Pérez, Carmen; School of Engineering and Sciences; Campus Monterrey; Lozano García, OmarThroughout history, legumes have been part of human consumption for their nutritional content and because is an easy crop to cultivate, it can grow in both cold and warm climates. One type of legumes are lentils, consumed worldwide. In Mexico, lentils are consumed by 70% of Mexican adults. Among the lentil varieties, green lentils stand out for having the highest polyphenol content, which makes them an excellent candidate for human consumption. However, the traditional way of cooking lentils requires prolonged times in boiling water. This causes a significant loss of the number of polyphenols present in lentils. Polyphenols have the ability to reduce the prevalence of suffering from chronic degenerative diseases, because they have antioxidants and anti-inflammatories properties. However, the chemical stability of polyphenols is compromised by different factors like the chemical structure, temperature, pH, isomerizations, enzymes, degradation, and oxidation, among others. This study subjected the green lentils to instant controlled pressure drop (DIC) and measured the polyphenol amount, flavonoids and antioxidant capacity 1,1 -diphenyl-2-picrylhydrazyl (DPPH) and Trolox equivalent antioxidant capacity (TEAC and DPPH), with 13 different treatments by varying pressure and time. The results showed that the polyphenols were the only parameter affected by DIC and the best conditions were less than 160 s and less than 0.1 MPa, and the best treatment was the DIC treatment 11, with 0.1 MPa for 135 s. Surprisingly, apparently new polyphenols appeared in the treated lentils due to the physical stress secondary to DIC, and in consequence the biosynthesis of polyphenols. After DIC, the best green lentil treatment was selected (DIC 11). The polyphenolic extract was obtained and nano encapsulated in poly lactic-co-glycolic acid (PLGA) using five different extract volumes (100, 250, 500, 750 and 1000 𝜇L). The nanoparticles were spherical in shape, with negative zeta potential charge (~ 20 mV), and all the syntheses produced particles, with average sizes ranging between 300 to 1100 nm. The polyphenol released was evaluated in PBS at pH 5.5 and 7.4. The release followed a triphasic controlled release, a lag phase of 24 h, a burst and diffusion phase from 24 h to 372 h, up to 15 days, and finally the saturation phase. The combination of the DIC technology as a pretreatment for green lentils and the nanoencapsulation in PLGA nanoparticles, improved the extraction and preserved the polyphenols profile of green lentils, on the other hand, nanoencapsulation protected the polyphenols and reached a controlled polyphenol release for up to 15 days.
- Design and Development of Conducting Polymer and Carbon Nanostructure based Efficient Thermoelectric Materials(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12-02) Ebrahimibagha, Dariush; Mallar, Ray; emimmayorquin; Aguirre Soto, Héctor Alán; Niladri, Banerjee; Gallo Villanueva, Roberto Carlos; School of Engineering and Sciences; Campus Monterrey; Datta, ShubhabrataThermoelectric materials present a promising renewable energy technology for directly converting thermal energy into electricity and vice versa. However, their practical application is hindered by low conversion efficiencies, quantified by the dimensionless figure of merit, 𝑍𝑇 = 𝑆 2 𝜎 𝑘 𝑇 , where 𝑆,𝜎, and 𝑘 are the Seebeck coefficient, electrical onductivity, and thermal conductivity, respectively. Achieving a high 𝑍𝑇 is challenging because enhancing one parameter often degrades the others. Various nanoscale strategies have been explored, yet a comprehensive framework for improving 𝑍𝑇 remains elusive. Recently, polymer-based nanocomposites, particularly carbon nanotubes (CNTs) dispersed in polyaniline (PANI), have gained attention due to their flexibility, non-toxicity, and processability, key traits for next-generation flexible electronic devices. Despite this potential, optimizing thermoelectric performance in PANI-CNT systems is complex, as it depends on numerous factors, including CNT dimensions, functionality, and PANI's doping and morphology. This research employs machine learning (ML) and genetic algorithms (GA) to model and optimize the thermoelectric properties of PANI-CNT nanocomposites. By analyzing structural and compositional variables—such as CNT length, diameter, type, and PANI morphology—we identified strategies that enhance electrical conductivity and the Seebeck coefficient while minimizing thermal conductivity. Our ML models revealed that selecting appropriate dopants for PANI and using single-walled CNT (SWCNT) improves overall thermoelectric performance. Multi-objective GA optimization further refined these findings, demonstrating that SWCNTs help reduce thermal conductivity and that CNT length plays a dual role: shorter CNTs decrease 𝑘, while longer ones enhance both 𝑆 and 𝜎. Experimental validation was performed by fabricating PANI-CNT nanocomposite pellets, but achieving high 𝑍𝑇 remained elusive due to limitations in dataset quality and the variability introduced by diverse synthesis techniques. The synthesis method influences PANI dimensionality (e.g., 0D, 1D, 2D) and the morphology of PANI-CNT composites (core-shell vs. dispersed), complicating performance consistency. While the experiments confirmed the general trend of model predictions, they highlighted the necessity of cleaner, more comprehensive datasets for future research. Ultimately, this study lays the groundwork for designing high-efficiency thermoelectric nanocomposites and outlines the next steps in developing more accurate predictive models and synthesis methods for improved thermoelectric performance.
- Modelling and Control Methodologies for Automated Systems Based on Regulation Control and Coloured Petri Nets(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12-02) Anguiano Gijón, Carlos Alberto; Vázquez Topete, Carlos Renato; emimmayorquin; Navarro Gutiérrez, Manuel; Navarro Díaz, Adrán; Mercado Rojas, José Guadalupe; School of Engineering and Sciences; Campus Monterrey; Ramírez Treviño, AntonioIndustry 4.0 and smart manufacturing have brought new interesting possibilities and chal-lenges to the industrial environment. One of these challenges is the large-scale automation of increasingly complex systems with minimal set-up time and flexibility, while allowing the in-tegration of components and systems from different manufacturers for production customiza-tion. To face this challenge, control approaches based on Discrete Event Systems (DES), such as Supervisory Control Theory (based on either, automata or Petri nets), Generalized Mutual Exclusions Constraints (GMEC) and Petri net-based Regulation Control, may provide con-venient solutions. However, few works have been reported in the literature for the case of complex systems and implementation in real plants. The latter opens up an important area of research opportunities. In this dissertation work, methodologies for modelling and control of automated systems based on the Regulation Control approach using interpreted Petri nets are studied. Using this approach, it is possible to capture the information of a system through its inputs and outputs, which allows to force sequences and generate more efficient controllers that can be directly translated to a Programmable Logic Controller (PLC). Through case studies, the effective-ness of these methodologies when implemented in more complex systems is demonstrated. Furthermore, the use of coloured Petri nets is proposed for the modelling of customized pro-duction systems. For this purpose, a new approach based on tensor arrays is introduced to express the colored Petri nets, allowing the use of algebraic techniques in the analysis of these systems.
- Botnet detection on twitter: a novel similarity-based clustering mechanism(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12) Samper Escalante, Luis Daniel; Monroy Borja, Raúl; emipsanchez; Castro Espinoza, Félix Agustín; González Mendoza, Miguel; School of Engineering and Sciences; Rectoría Tec de Monterrey; Loyola González, OctavioBotnet detection on Twitter represents a critical yet under-explored research problem,as botnets programmed with malicious intent threaten the platform’s security and credibility. Although Twitter has implemented mitigation strategies, such as imposing restrictions andbans, these measures remain insufficient due to botnets’ rapid creation and expansion. Existing solutions proposed by researchers for manual and automated botnet detection typically rely on individual metrics commonly used for detecting bots. However, these approaches lack the necessary group-oriented analysis and metrics critical for effectively identifying botnets of varying sizes and objectives. To address this issue, we have developed an innovative botnet detection mechanism based on similarity, which significantly enhances the detection rate of botnets on Twitter. Each bot, regardless of its complexity, leaves detectable traces of automation in its creation, behavior, or interactions with other accounts. By characterizing these traces, we can establish relationships between bots, enabling effective botnet detection. Our mechanism constructs a regression model to quantify the similarity between bots, leveraging features from user data, tweet patterns, and social interactions on the platform. Then, it uses this similarity measure to build a distance matrix, enabling the formation of groups with shared attributes, connections, and objectives through clustering methods. Our botnet detection mechanism achieved extraordinary success, evidenced by high scores on external Clustering Validation Indices (CVIs) and the Area under the ROC Curve (AUC) compared to existing solutions from the literature. Furthermore, the mechanism proved effective when confronted with unknown botnets with varied objectives. Our experimental findings suggest that this work is well-positioned to strengthen future botnet detection mechanisms, having shown the value of incorporating social interaction features. This integration offers a strategic advantage in the ongoing arms race against botmasters and their malicious objectives. Additionally, our mechanism consistently outperforms other approaches across various metrics, configurations, and algorithms, underscoring its effectiveness and adaptability in different detection scenarios.
- Wounding stress and UVB radiation for increasing anti-obesogenic compounds in raw vegetables – a practical approach(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12) Gastélum Estrada Alejandro; Reza Zaldívar, Edwin Estefan; emipsanchez; Marilena Antunes Ricardo; Reza Zaldívar, Edwin Estefan; Canales Aguirre, Alejandro Arturo; Benavides Lozano, Jorge; School of Engineering and Sciences; Campus GuadalajaraThe rise of uncommunicable diseases such as obesity, type 2 diabetes, and cardiovascular diseases has urged the development of innovative approaches to improve public health, particularly through dietary interventions. Fruits and vegetables are widely recognized as important sources of bioactive compounds , including phenolics, carotenoids, and flavonoids, which contribute to their health-promoting properties. Despite the known benefits, global vegetable consumption remains below recommended levels, leaving many populations at risk for these diet-related diseases. This work explores the potential of two postharvest abiotic stresses, wounding stress and ultraviolet B (UVB) radiation, as techniques for increasing the bioactive compound content of commonly consumed vegetables. This study propases adapting these stress techniques (wounding stress and UVB radiation) for domestic environments. These adaptations would give consumers a practica! Do-lt Yourself (DIY) approach to improving the bioactive content of their vegetable consumption, making it possible to obtain greater health benefits from smaller portions. Three widely consumed vegetables (carrots, broccol,i and lettuce) were chosen for this study due to their high consumption rates and nutraceutical potential. The first part of the research involved applying wounding stress to carrots through shredding, followed by storage at 15ºC for 48 hours to allow for phenolic accumulation . The stressed carrots were then used to prepare a biofortified juice, blended with orange juice, broccoli sprouts, and pasteurized. Physicochemical attributes and bioactive compound content were evaluated weekly in juice during a storage period of 28 days. Bioactivity was assessed in vitro at daysO and 28 after preparation. Results showed that the stressed carrots significantly increased total phenolic content, chlorogenic acid, and glucosinolates. The biofartified juice displayed enhanced antioxidant and anti-inflammatory properties, which were preserved throughout storage. Far the UVB radiation stress study, a chamber was developed to treat vegetables, including carrots, broccoli, and lettuce. The chamber configuration and exposure conditions (time and intensity) were optimized to deliver the appropriate UVB dose to maximize phenolic compounds accumulation. The treated vegetables were then assessed far their bioactive compound content by chromatography analysis ; antioxidant capacity, anti-inflammatory potential, and anti-obesogenic potential were evaluated in vitro. Results indicated significant increases in phenolic compounds in the three evaluated vegetables and glucosinolates far broccoli in the UVB-treated vegetables, with enhanced antioxidant and anti-inflammatory properties. The findings of this research confirm the efficacy of wounding stress and UVB radiation in increasing the bioactive compound content of vegetables, demonstrating that these techniques can be successfully applied in domestic and industrial contexts. The development of a UVB chamber far home use also represents a significant innovation, offering consumers a practica! tool far enhancing the health benefits of their vegetables. This research opens a new opportunity far improving diet quality through scalable and affordable techniques. The DIY approach offers an accessible strategy far individuals to increase their vegetable consumption's health impact. The perspectives of this work suggest broader applications in restaurants, schools, and other faod service environments, where these techniques could be implemented to improve the nutritional quality of meals served, which potentially clase the gap between low vegetable intake and bioactive compounds consumption needed to reduce the risk of uncommunicable diseases.
- Exact and heuristic approaches for solving coverage and agricultural problems(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12) Vicencio Medina, Salvador de Jesús; Ríos Solís, Yasmín Águeda; emimmayorquin; Cárdenas Barrón, Leopoldo Ediro; School of Engineering and Sciences; Cid García, NéstorIn this dissertation, two problems have been studied. The first problem, the maximal covering location problem with accessibility indicators and mobile units, belongs to the humanitarian logistic area. The second problem, the orthogonal site-specific management zone, belongs to the precision agriculture area. Both issues have been addressed through Operation Research techniques, which have a general purpose and can be used for various problems. The objective of the maximal covering location problem with accessibility indicators and mobile units is to allocate the COVID-19 tests to hospitals in Mexico to maximize the coverage, the number of opportunities, the service network, and other accessibility measures. To solve the maximal covering location problem with accessibility indicators and mobile units, a mathematical model that incorporates different accessibility measures and mobile units has been proposed. The mathematical model can solve small and medium instances in a short computational time. A matheursitic that combines an Estimation of Distribution Algorithm with a version parameterized of the proposed model has also been developed to solve medium and large instances. The computational results show that incorporating the mobile units with the accessibility measures considered has a significant improvement compared with the literature approaches. The orthogonal site-specific management zone problem aims to determine the minimum number of site-specific management zones that fulfill a homogeneity level measured through the relative variance. The zones must also be orthogonal since these shapes make it practical to delineate them for traditional agricultural machinery. An approximate and exact approach has been proposed to solve the orthogonal site-specific management zone problem.The approximate approach consists of a metaheuristic known as the Estimation of Distribution Algorithm. It uses a special decoder based on disjoint sets and a new reactive fitness function to provide high-quality solutions in short computational times. The results improve the solutions’ quality and computational times presented in the literature. Additionally, a new data set of instances has been proposed due to the results and times obtained with this approach. Using this new data set, the algorithm proposed continues to be fast and obtain quality solutions. Two mathematical programming formulations and one constraint programming formula-tion have been proposed for the exact approach. The mathematical programming formulations yield three cutting-plane algorithms. The formulations proposed obtain high-quality solutions for small and medium instances in short computational times. Besides, the formulations mentioned consider orthogonal management zones and the relative variance as constraints. To our knowledge, only heuristic methods have addressed this problem. Thus, the formulations presented in this work are the first in the literature to solve the orthogonal site-specific manage-ment zone problem. The computational results show that the formulations proposed obtain optimality for small and medium instances. Besides, these results make it possible to compare and validate the results obtained through the heuristics methods present in the literature.
- Advanced modeling techniques in electric vehicles for battery sizing and Vertical Dynamic Control with CARSIM® and ADAMS(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12) Drivet González, Aline Raquel Lily; Cespi, Riccardo; emipsanchez; Vargas Martínez, Adriana; Lozoya Santos, Jorge de Jesús; School of Engineering and Sciences; Campus Monterrey; Tudón Martínez, Juan CarlosThis thesis addresses the rapidly accelerating shift from internal combustion engine vehicles to electric vehicles (EVs), a transition driven not only by market demands but also by the urgent need to mitigate climate change. As electrification reshapes the automotive landscape, the importance of advanced modeling techniques are essential to accelerate the adoption of EV technologies, ensuring competitiveness, and addressing environmental urgency. This research begins with a review of vehicle dynamics changes, highlighting the challenges and opportunities introduced by this swift transition to EV technology. The first contribution of this thesis is the application of modeling and simulation techniques using CARSIM®where real-world telemetry is used to optimize EV battery performance and battery sizing. This optimization focuses on maximizing efficiency while maintaining safety and reliability. The second contribution is the development of a model for EV suspension systems using ADAMS®which can be a platform to test critical dynamic behavior of EVs under various conditions. Together, these contributions advance the design and performance of electric vehicles, introducing advanced modeling tools to accelerate development processes, speeding design processes, and addressing the urgent challenges of vehicle electrification in the context of climate change. As a result of the research presented in this thesis, which includes methodologies for battery pack design and the modeling and control of active suspension systems for electric vehicles, two journal articles have been published, and four additional articles have been presented in conference proceedings, contributing significantly to the academic discourse in these areas.
