Optimization models applied to the location of charging stations for electric vehicles
Authorship
A.C.P.
Master in Statistical Techniques
A.C.P.
Master in Statistical Techniques
Defense date
09.05.2025 10:15
09.05.2025 10:15
Summary
This work presents a review of optimization models applied to the location of charging stations for electric vehicles, a key aspect in the expansion of this infrastructure to promote the use of electric vehicles as a measure to reduce emissions and support the transition towards more sustainable mobility. We analyze the main existing approaches, classified into two broad categories according to how demand is modeled: node-based models, typically applied in urban environments, and flow-based models, where demand is represented by origin-destination trips and is more suitable for larger geographic areas. The basic formulations are presented, along with the main ways of incorporating other relevant aspects of the problem, such as capacity constraints or the specific characteristics of the transport system. In addition, we review the most common solution methods, usually based on mixed-integer linear programming. Finally, the review is complemented with a practical case study applied to Galicia, which illustrates the application of these models in a context close to reality and allows for a comparison of their performance and results.
This work presents a review of optimization models applied to the location of charging stations for electric vehicles, a key aspect in the expansion of this infrastructure to promote the use of electric vehicles as a measure to reduce emissions and support the transition towards more sustainable mobility. We analyze the main existing approaches, classified into two broad categories according to how demand is modeled: node-based models, typically applied in urban environments, and flow-based models, where demand is represented by origin-destination trips and is more suitable for larger geographic areas. The basic formulations are presented, along with the main ways of incorporating other relevant aspects of the problem, such as capacity constraints or the specific characteristics of the transport system. In addition, we review the most common solution methods, usually based on mixed-integer linear programming. Finally, the review is complemented with a practical case study applied to Galicia, which illustrates the application of these models in a context close to reality and allows for a comparison of their performance and results.
Direction
Carpente Rodríguez, Maria luisa (Tutorships)
Carpente Rodríguez, Maria luisa (Tutorships)
Court
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Vidal Puga, Juan José (Chairman)
Oviedo de la Fuente, Manuel (Secretary)
PATEIRO LOPEZ, BEATRIZ (Member)
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Vidal Puga, Juan José (Chairman)
Oviedo de la Fuente, Manuel (Secretary)
PATEIRO LOPEZ, BEATRIZ (Member)
Analysis of the effect of e-appointment in the Cardiology Service of the Santiago de Compostela and Barbanza Health Area
Authorship
I.I.G.
Master in Statistical Techniques
I.I.G.
Master in Statistical Techniques
Defense date
09.03.2025 09:00
09.03.2025 09:00
Summary
In this work, we will analyze the effect of the introduction of electronic appointment (usually known as e-appointment) in the Cardiology Service for patients associated with the Santiago de Compostela and Barbanza Health Area. To do this, we will use generalized additive models (commonly known as GAM), which allow us to model non-parametric relationships between explanatory variables and the several selected response variables. First, we will study the impact of the implementation of e-appointment in order to improve the accessibility in the Galician healthcare system. Specifically, we will focus on key aspects of clinical management, such as hospital admissions due to cardiovascular causes, delay time for access to the Cardiology Service, the number of emergency visits or the number of appointments related to pathologies managed by the Cardiology Service. Next, taking into account the presence of censored data in the considered dataset, we will analyze the influence of e-appointment on patient life time. For this purpose, we will once again use GAMs, incorporating Kaplan-Meier weights in order to adapt the methodology to the context of censored data.
In this work, we will analyze the effect of the introduction of electronic appointment (usually known as e-appointment) in the Cardiology Service for patients associated with the Santiago de Compostela and Barbanza Health Area. To do this, we will use generalized additive models (commonly known as GAM), which allow us to model non-parametric relationships between explanatory variables and the several selected response variables. First, we will study the impact of the implementation of e-appointment in order to improve the accessibility in the Galician healthcare system. Specifically, we will focus on key aspects of clinical management, such as hospital admissions due to cardiovascular causes, delay time for access to the Cardiology Service, the number of emergency visits or the number of appointments related to pathologies managed by the Cardiology Service. Next, taking into account the presence of censored data in the considered dataset, we will analyze the influence of e-appointment on patient life time. For this purpose, we will once again use GAMs, incorporating Kaplan-Meier weights in order to adapt the methodology to the context of censored data.
Direction
CONDE AMBOAGE, MERCEDES (Tutorships)
CONDE AMBOAGE, MERCEDES (Tutorships)
Court
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Bergantiños Cid, Gustavo (Chairman)
GINZO VILLAMAYOR, MARIA JOSE (Secretary)
Darriba López, Diego (Member)
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Bergantiños Cid, Gustavo (Chairman)
GINZO VILLAMAYOR, MARIA JOSE (Secretary)
Darriba López, Diego (Member)
Uncertainty quantification in anomaly detection models
Authorship
C.M.O.
Master in Statistical Techniques
C.M.O.
Master in Statistical Techniques
Defense date
09.05.2025 11:00
09.05.2025 11:00
Summary
This Master's Thesis presents a study on the application of uncertainty quantification methodologies to anomaly detection algorithms in unsupervised contexts, applied to a case of industrial security. Anomaly detection methods are not free from errors, so measuring the uncertainty of these models and their predictions is crucial, especially in critical contexts such as cybersecurity. The methodology is evaluated in two scenarios: a synthetic one, which facilitates graphical representation of its functioning, and a more realistic one based on network traffic data generated in a laboratory environment.
This Master's Thesis presents a study on the application of uncertainty quantification methodologies to anomaly detection algorithms in unsupervised contexts, applied to a case of industrial security. Anomaly detection methods are not free from errors, so measuring the uncertainty of these models and their predictions is crucial, especially in critical contexts such as cybersecurity. The methodology is evaluated in two scenarios: a synthetic one, which facilitates graphical representation of its functioning, and a more realistic one based on network traffic data generated in a laboratory environment.
Direction
SESTELO PEREZ, MARTA (Tutorships)
SESTELO PEREZ, MARTA (Tutorships)
Court
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Vidal Puga, Juan José (Chairman)
Oviedo de la Fuente, Manuel (Secretary)
PATEIRO LOPEZ, BEATRIZ (Member)
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Vidal Puga, Juan José (Chairman)
Oviedo de la Fuente, Manuel (Secretary)
PATEIRO LOPEZ, BEATRIZ (Member)
Study of the bed needs for tumor pathologies at the level of Autonomous Communities.
Authorship
C.A.P.V.
Master in Statistical Techniques
C.A.P.V.
Master in Statistical Techniques
Defense date
09.05.2025 16:45
09.05.2025 16:45
Summary
In the field of hospital management, there is a need for an adequate allocation of beds in order to provide quality medical care for patients with lung cancer at the level of the autonomous communities. This work seeks to solve this problem using Operations Research tools, formulating a robust deterministic model that addresses all the stated needs, comparing it with different solvers through examples with real data. The conclusions of this study ensure that the choice of model helps us meet the healthcare needs in all hospitals of the autonomous communities of Spain.
In the field of hospital management, there is a need for an adequate allocation of beds in order to provide quality medical care for patients with lung cancer at the level of the autonomous communities. This work seeks to solve this problem using Operations Research tools, formulating a robust deterministic model that addresses all the stated needs, comparing it with different solvers through examples with real data. The conclusions of this study ensure that the choice of model helps us meet the healthcare needs in all hospitals of the autonomous communities of Spain.
Direction
GINZO VILLAMAYOR, MARIA JOSE (Tutorships)
SAAVEDRA NIEVES, ALEJANDRO (Co-tutorships)
GINZO VILLAMAYOR, MARIA JOSE (Tutorships)
SAAVEDRA NIEVES, ALEJANDRO (Co-tutorships)
Court
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Vidal Puga, Juan José (Chairman)
Oviedo de la Fuente, Manuel (Secretary)
PATEIRO LOPEZ, BEATRIZ (Member)
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Vidal Puga, Juan José (Chairman)
Oviedo de la Fuente, Manuel (Secretary)
PATEIRO LOPEZ, BEATRIZ (Member)
Schedule Optimization
Authorship
S.R.G.
Master in Statistical Techniques
S.R.G.
Master in Statistical Techniques
Defense date
09.03.2025 11:45
09.03.2025 11:45
Summary
This work proposes a mixed integer linear optimization model for scheduling in the retail sector, building upon and expanding the classical Nursing Personnel Scheduling Problem. Its objective is to assign shifts efficiently while respecting labor constraints, individual preferences, specific skills, and the operational needs of the store. The solution is integrated into a broader personnel management process, structured in phases ranging from constraint preprocessing to robustness-based filtering. An integer linear programming model is presented for weekly planning, complemented by algorithms that enable scalability and reduce computational load. The results support the feasibility of the proposed approach and pave the way toward more flexible systems tailored to the commercial context.
This work proposes a mixed integer linear optimization model for scheduling in the retail sector, building upon and expanding the classical Nursing Personnel Scheduling Problem. Its objective is to assign shifts efficiently while respecting labor constraints, individual preferences, specific skills, and the operational needs of the store. The solution is integrated into a broader personnel management process, structured in phases ranging from constraint preprocessing to robustness-based filtering. An integer linear programming model is presented for weekly planning, complemented by algorithms that enable scalability and reduce computational load. The results support the feasibility of the proposed approach and pave the way toward more flexible systems tailored to the commercial context.
Direction
GONZALEZ RUEDA, ANGEL MANUEL (Tutorships)
GONZALEZ RUEDA, ANGEL MANUEL (Tutorships)
Court
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Bergantiños Cid, Gustavo (Chairman)
GINZO VILLAMAYOR, MARIA JOSE (Secretary)
Darriba López, Diego (Member)
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Bergantiños Cid, Gustavo (Chairman)
GINZO VILLAMAYOR, MARIA JOSE (Secretary)
Darriba López, Diego (Member)
Non-parametric curve estimation for biased data
Authorship
N.S.M.
Master in Statistical Techniques
N.S.M.
Master in Statistical Techniques
Defense date
09.05.2025 13:00
09.05.2025 13:00
Summary
In general terms, it is assumed that a sample accurately reflects the fundamental characteristics of the population it represents. However, in reality, many samples are generated within a context where the probability of including an observation in the sample is conditioned by its own value or other relevant variables, giving rise to what is known as biased data. The main objective of this work is to adapt classical techniques of statistical inference to address the complexities arising from biased data. We focus particularly on the estimation of general interest curves, such as the density function and the distribution function. The theoretical content is complemented by corresponding implementations in R. To illustrate more specifically the application of these techniques, actual data related to the size of shrubs of the species Cercocarpus montanus in a former limestone quarry located east of Laramie, Wyoming, are used.
In general terms, it is assumed that a sample accurately reflects the fundamental characteristics of the population it represents. However, in reality, many samples are generated within a context where the probability of including an observation in the sample is conditioned by its own value or other relevant variables, giving rise to what is known as biased data. The main objective of this work is to adapt classical techniques of statistical inference to address the complexities arising from biased data. We focus particularly on the estimation of general interest curves, such as the density function and the distribution function. The theoretical content is complemented by corresponding implementations in R. To illustrate more specifically the application of these techniques, actual data related to the size of shrubs of the species Cercocarpus montanus in a former limestone quarry located east of Laramie, Wyoming, are used.
Direction
BORRAJO GARCIA, MARIA ISABEL (Tutorships)
CONDE AMBOAGE, MERCEDES (Co-tutorships)
BORRAJO GARCIA, MARIA ISABEL (Tutorships)
CONDE AMBOAGE, MERCEDES (Co-tutorships)
Court
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Vidal Puga, Juan José (Chairman)
Oviedo de la Fuente, Manuel (Secretary)
PATEIRO LOPEZ, BEATRIZ (Member)
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Vidal Puga, Juan José (Chairman)
Oviedo de la Fuente, Manuel (Secretary)
PATEIRO LOPEZ, BEATRIZ (Member)
The impact of climate change on the retail sector’s demand
Authorship
T.T.
Master in Statistical Techniques
T.T.
Master in Statistical Techniques
Defense date
09.05.2025 09:30
09.05.2025 09:30
Summary
This Master's Thesis presents a detailed analysis of the sales and returns of the partner company INDITEX. The study includes an exhaustive exploratory analysis aimed at understanding the impact of weather conditions on customer consumption patterns, as well as the development of several predictive models focused on forecasting the number of units sold and returned.
This Master's Thesis presents a detailed analysis of the sales and returns of the partner company INDITEX. The study includes an exhaustive exploratory analysis aimed at understanding the impact of weather conditions on customer consumption patterns, as well as the development of several predictive models focused on forecasting the number of units sold and returned.
Direction
AMEIJEIRAS ALONSO, JOSE (Tutorships)
SAAVEDRA NIEVES, PAULA (Co-tutorships)
AMEIJEIRAS ALONSO, JOSE (Tutorships)
SAAVEDRA NIEVES, PAULA (Co-tutorships)
Court
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Vidal Puga, Juan José (Chairman)
Oviedo de la Fuente, Manuel (Secretary)
PATEIRO LOPEZ, BEATRIZ (Member)
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Vidal Puga, Juan José (Chairman)
Oviedo de la Fuente, Manuel (Secretary)
PATEIRO LOPEZ, BEATRIZ (Member)
Optimisation problems: Machine Learning vs. traditional methods
Authorship
M.V.A.
Master in Statistical Techniques
M.V.A.
Master in Statistical Techniques
Defense date
09.03.2025 13:45
09.03.2025 13:45
Summary
From the SDG consultancy, the objective of this project was presented as a comparison in terms of economic efficiency between two different approaches for defining the values of a MIN and MAX model. Every inventory management process requires both a demand forecasting process and a resource optimization process. The objective of this project is to determine whether the classical approach (forecasting demand and then optimizing) is substantially better than a new approach, which performs an optimization of the MIN and MAX values on historical data and uses these values to make the forecast.
From the SDG consultancy, the objective of this project was presented as a comparison in terms of economic efficiency between two different approaches for defining the values of a MIN and MAX model. Every inventory management process requires both a demand forecasting process and a resource optimization process. The objective of this project is to determine whether the classical approach (forecasting demand and then optimizing) is substantially better than a new approach, which performs an optimization of the MIN and MAX values on historical data and uses these values to make the forecast.
Direction
GONZALEZ DIAZ, JULIO (Tutorships)
PATEIRO LOPEZ, BEATRIZ (Co-tutorships)
GONZALEZ DIAZ, JULIO (Tutorships)
PATEIRO LOPEZ, BEATRIZ (Co-tutorships)
Court
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Bergantiños Cid, Gustavo (Chairman)
GINZO VILLAMAYOR, MARIA JOSE (Secretary)
Darriba López, Diego (Member)
AMEIJEIRAS ALONSO, JOSE (Coordinator)
Bergantiños Cid, Gustavo (Chairman)
GINZO VILLAMAYOR, MARIA JOSE (Secretary)
Darriba López, Diego (Member)