Embedded Files
The Industrial Engineering Seminar is a bi-weekly gathering designed to unite faculty, students and practitioners with a keen interest in Industrial Engineering research, especially in the realms of Operations and Organisational Management.
Our sessions are primarily tailored for students. We aim to unveil the dynamic connections between research, academia and the practical world, demonstrating the profound impacts that engineering tools can have on society, government and businesses. Each seminar commences with a casual conversation with our esteemed speakers. This segment is dedicated to exploring the potential impacts of their research, delving into their personal motivations and interests, and uncovering their favoured areas of investigation.
Meetings typically take place in the Auditorium of the Industrial Engineering Department, located on the 3rd floor of the south building at the University of Santiago (USACH). You can enter directly from the internal street.
Keep informed by subscribing to our Google Calendar and mailing list.
Organisers: Fernando Garcia & Sebastián Dávila-Gálvez.
Former Organisers: Franco Quezada, Fabián Fuentes González & Jorge Zamorano, Francisco Jara & Guido Lagos
Do not hesitate to email us if you are interested in giving a talk! - seminario.iseindustrias@usach.cl
Keep informed by subscribing to our Google Calendar and mailing list.
Organisers: Fernando Garcia & Sebastián Dávila-Gálvez.
Former Organisers: Franco Quezada, Fabián Fuentes González & Jorge Zamorano, Francisco Jara & Guido Lagos
Do not hesitate to email us if you are interested in giving a talk! - seminario.iseindustrias@usach.cl
Upcoming talks
Upcoming talks
Armin Lüer-Villagra
Armin Lüer-Villagra
Profesor investigador, UNAB
Miércoles, 02 de julio, 11:30am - 12:45am.
Cowork, Departamento de Ingeniería Industrial. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: A stochastic hub location and protection problem
Abstract: This research addresses the design of resilient hub-and-spoke networks under uncertainty, where a company must decide on the location of its hubs and which ones to protect against potential attacks. Unlike traditional models with deterministic strategies, this work considers stochastic attack scenarios with unknown interdiction plans.
The objective is to minimize pre-attack routing costs and expected disruption costs due to demand loss, while anticipating adversarial actions that aim to maximize disruption. The proposed model leverages a two-level stochastic programming approach and incorporates acceleration techniques such as variable fixing and scenario-based decomposition. The solution methods are implemented using the CPLEX C++ API, yielding optimal or near-optimal results within reasonable computational times.
Constanza Contreras
Co-founder de X-Academy, y PhD(c) en sistemas de ingeniería, UChileConstanza Contreras
Miércoles, 25 de junio, 11:30am - 12:45am.
Auditorio, Departamento de Ingeniería Industrial. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: Challenges in the Management of Waiting Times for Elective Surgeries: National and International Perspectives
Abstract: Prolonged waiting times for elective surgeries have become a critical issue in Chile’s public healthcare system, which serves nearly 80% of the population. The COVID-19 pandemic has exacerbated this challenge, revealing systemic inefficiencies and resource constraints. Despite recent reforms, such as the establishment of Regional Resolution Centers (CRR), waiting lists continue to grow, impacting patient well-being and care quality. This talk explores the structural and operational causes behind surgical delays, compares Chile’s situation with international benchmarks, and demonstrates how operations research tools—such as optimization, simulation, and data analytics—can support more effective and equitable healthcare delivery and inform public policy.
Iván Derpich- Universidad de Sanatigo de Chile
Profesor asistenteIván Derpich- Universidad de Sanatigo de Chile
Miércoles, 2185 de junio, 11:30am - 12:45am.
Auditorio, Departamento de Ingeniería Industrial. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: Un modelo teórico para el cálculo de la huyella de carbono en pueros basado en relajación lagrangiana
Abstract: Los puertos desempeñan un papel estratégico en el desarrollo de los países, y la transición hacia energías limpias está transformando su operación. En los países emergentes, se hace cada vez más necesario incorporar maquinaria eléctrica en las terminales portuarias, siguiendo los estándares ambientales ya adoptados por naciones desarrolladas. Esta charla presenta un análisis comparativo de las políticas y criterios energéticos en varios puertos de un país emergente y un puerto europeo. Se optimiza el diseño del patio de carga y se estiman las emisiones de gases de efecto invernadero (GEI) bajo operación diésel y eléctrica, cuantificando así los beneficios ambientales de la electrificación. Se discuten brechas estratégicas y operativas en la gestión energética entre países emergentes y desarrollados, y se reflexiona sobre cómo estas iniciativas pueden contribuir a la sostenibilidad urbana.
Eduardo Zúñiga Leyton- U. Diego Portales - Chile
Profesor asistenteEduardo Zúñiga Leyton- U. Diego Portales - Chile
Miércoles, 14 de mayo, 11:30am - 12:45am.
Auditorio, Departamento de Ingeniería Industrial. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: El rol del timing en las interacciones estratégicas
Abstract: ¿Cómo afecta el orden en que los agentes toman decisiones estratégicas? Esta presentación aborda esa pregunta a través del concepto de "endogenous timing'' en Teoría de Juegos. Partiendo desde algunos conceptos básicos, exploraremos cómo el orden de las decisiones, que usualmente se impone de manera exógena, puede surgir como parte del equilibrio.
Revisaremos dos esquemas clásicos para "endogenizar'' el timing de las decisiones: el Game with Observable Delay (GOD) y el Game with Action Commitment (GAC). Veremos su estructura general, los compararemos en distintos contextos, y discutiremos en qué casos predicen resultados diferentes. Finalmente, veremos cómo se aplican estos esquemas a tres escenarios relevantes: competencia entre firmas, juegos de seguridad y provisión de bienes públicos.
Gonzalo Méndez Vogel- UNAB - Chile
Profesor asistenteGonzalo Méndez Vogel- UNAB - Chile
Miércoles, 14 de mayo, 11:30am - 12:45am.
Auditorio, Departamento de Ingeniería Industrial. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: Localización de Puntos Limpios para Maximizar el Reciclaje: Un Caso de Estudio en Estación Central
Abstract: Ubicar contenedores de reciclaje en una ciudad no es tan simple como ponerlos cerca de las personas. Aunque la proximidad ayuda, no siempre garantiza que la gente recicle: algunos no lo hacen por falta de interés o tiempo, otros eligen qué contenedor usar según factores como la iluminación de la calle o el entorno, y además, muchas veces reciclan incluso estando más lejos. A esto se suma que los usuarios no saben si un contenedor está lleno hasta llegar al lugar, lo que puede provocar desbordes y pérdida de material reciclable. Para enfrentar estos desafíos, este estudio propone un modelo probabilístico basado en el logit multinomial, que representa de manera más realista las decisiones de los usuarios. Además, se incorporan restricciones flexibles que modelan el comportamiento frente al desbordamiento de los contenedores, con el objetivo de maximizar la cantidad de material reciclado y minimizar los residuos no recogidos. El problema de optimización planteado se formula como un modelo entero no lineal, el cual es reformulado de manera lineal y resuelto mediante técnicas de ramificación y corte (branch-and-cut). Finalmente, el modelo se aplica en un caso de estudio en la comuna de Estación Central, Santiago de Chile, mostrando patrones de ubicación consistentes con los comportamientos observados en la vida real.
Natalia Jorquera - ENSTA - Paris
Phd(c) - Institut Prof de Paris, France.Natalia Jorquera - ENSTA - Paris
Miércoles, 16 de abril, 11:30am - 12:45am.
Auditorio, Departamento de Ingeniería Industrial. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: Fair electricity supply planning for collective self-consumption
Abstract: This presentation explores fair electricity supply planning in collective self-consumption communities, focusing on shared distributed energy sources and a common electricity storage system. We propose a methodology to ensure a fair allocation of these shared resources. Our approach formulates the problem as a mixed-integer linear programming (MILP) model, later transformed into a linear programming (LP) model. We introduce different fairness criteria, such as proportional allocation and max-min fairness, and analyze their dominance properties. Through numerical experiments based on a real case study, we demonstrate the impact of these fairness models on green electricity usage, resource allocation, and participant costs.
Past talks 2024
Past talks 2024
Ángela Flores - Universidad de Chile
Phd - Departamento de Ingeniería Eléctrica - Universidad de Chile.Ángela Flores - Universidad de Chile
Lunes, 18 de Noviembre, 11:30am - 12:45am.
Auditorio, Departamento de Ingeniería Industrial. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: Power system planning under multi-scale climate-induced uncertainty
Abstract: Power systems are facing increasing pressure from climate change. The rise in global temperatures, more variable precipitation patterns, and the increase in frequency and severity of extreme weather events pose substantial risks for power systems across multiple time scales. In this talk, a comprehensive power system planning model is introduced to anticipate both short- and long-term uncertainties by leveraging multi-scale stochastic programming. Furthermore, a novel approach is presented for constructing multi-scale scenario trees to represent climate-induced uncertainty, using historical weather data and future climate projections. Finally, suitable decomposition methods for capacity planning under uncertainty and possible enhancements are explored.
Jornada de Finanzas Sostenibles y Verdes USACH 2024
Jornada de Finanzas Sostenibles y Verdes USACH 2024
Miércoles, 04 de septiembre, 09:00am - 13:00am.
Auditorio, Departamento de Ingeniería Industrial. Av. Víctor Jara 3769, Metro U. de Santiago.
Las finanzas sostenibles se refieren a cómo los actores del sistema financiero incorporan variables Ambientales, Sociales y de Gobernanza (ASG o ESG en inglés) en sus operaciones y decisiones, buscando consolidar una economía que beneficie a diversos grupos de interés. Por otro lado, las finanzas verdes, un subconjunto de las finanzas sostenibles, se enfocan en apoyar proyectos e iniciativas que contribuyan a mejorar el medio ambiente.
En esta jornada, abordaremos temas como el rol de la academia en finanzas sostenibles, el desempeño financiero de empresas comprometidas con los criterios ASG, avances regulatorios en Chile respecto a finanzas verdes, y el financiamiento de proyectos complejos de energía limpia y combate al cambio climático. Contaremos con la participación de especialistas de instituciones relevantes del país, incluyendo la Fundación Chile y el Ministerio de Hacienda, así como investigadores y académicos del Departamento de Ingeniería Industrial de nuestra universidad.
Esta jornada está dirigida a toda la comunidad académica, incluyendo profesores, estudiantes, personal administrativo, y a cualquier persona interesada en temas de sostenibilidad, medio ambiente y finanzas. Se espera que quienes asistan estén interesados en conocer más sobre estos conceptos y cómo la academia, el sector público y privado, y la sociedad en su conjunto pueden acelerar la transición hacia una economía más sostenible.
Marc Juanpera - Universitat Politècnica de Catalunya
Phd - Escuela Técnica Superior de Ingeniería Industrial de Barcelona - Universitat Politècnica de Catalunya, España.Marc Juanpera - Universitat Politècnica de Catalunya
Lunes, 02 de sepiembre, 11:30am - 12:45am.
Salón de honor, Departamento de Ingeniería Industrial, 3er piso edificio sur. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: Investigación operativa para la optimización de la distribución social de alimentos a familias vulnerables.
Bio: Marc Juanpera es Doctor Ingeniero Industrial por la Universitat Politècnica de Catalunya (UPC) y Profesor Lector (equivalente Assistant Professor) en el Departamento de Organización de Empresas de la UPC desde Septiembre de 2023. Se ha especializado en el desarrollo y aplicación de métodos cuantitativos de investigación operativa (principalmente programación lineal entera y mixta, algoritmos heurísticos y métodos multicriterio) para resolver problemas relacionados con la planificación de la producción, el acceso a energía en zonas rurales y, más recientemente, la optimización de la distribución social de alimentos. En esta última línea, está liderando la colaboración con entidades locales (dispensadoras de alimentos en Barcelona) e internacionales (Programa Mundial de Alimentos) para mejorar el acceso a una alimentación nutritiva y equitativa para familias en situación de vulnerabilidad.
Felipe Rivera - Universidad de Navarra
Phd - School of Economics and Business - University of Navarra, Spain.Felipe Rivera - Universidad de Navarra
Wednesday, July 24th, 11:30am - 12:45am.
Salón de honor, Departamento de Ingeniería Industrial, 3er piso edificio sur. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: Corporate Purpose: An Essential Lever for Sustainability and Business Competitivenessgeneration of renewable energy.
Abstract: The orientation towards profit maximization has contributed to the social and environmental issues we face today, as decisions are solely based on this principle, focusing exclusively on shareholders, leaving aside other important stakeholders (e.g., consumers, suppliers, regulators, and society at large). In response to this scenario, there is a growing interest in corporate purpose, representing a new approach in business management. Its influence on organizational commitment, cohesion, performance, and resilience is acknowledged. In this study, a conceptual framework is presented to understand a purpose-driven company, and its results are analyzed in economic terms. Additionally, exploration is made into how corporate governance can facilitate the adoption of a purpose, from new legal forms to the commitment of executives with stakeholders. Addressing this challenge involves investigating the implementation of purpose within companies and its effects in various areas of the organization. This study contributes to understanding how purpose can be effectively integrated into business strategy, promoting both economic and social well-being.
Ruiwen Liao (Université Paris-Saclay)
Phd(c) - LISN - Paris, France.Ruiwen Liao (Université Paris-Saclay)
Wednesday, June 19th, 11:30am - 12:45am.
Salón de honor, Departamento de Ingeniería Industrial, 3er piso edificio sur. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: A multi-stage stochastic programming model for lot-sizing with onsite generation of renewable energy.
Abstract: One way to achieve energy efficiency in manufacturing is to equipe plants with on-site renewable energy generation systems to partially power industrial processes. However, renewable energy sources are highly intermittent and their availability is difficult to predict accurately. Therefore, we study an integrated industrial production planning and energy supply problem under uncertain renewable energy availability. The resulting production and energy supply planning can be seen as a multi-stage decision process where some decisions are made at the beginning of the planning horizon whereas the others are postponed to later decision stages when more information on the uncertain parameters are revealed. At the beginning of the planning horizon, we build a setup and startup plan for a proportional lot-sizing and scheduling problem in a single-machine multi-item setting. Computational experiments will be presented to show the practical efficiency of the proposed approach.
Natalia Jorquera (ENSTA - Paris)
Phd(c) - ENSTA - Paris, France.Natalia Jorquera (ENSTA - Paris)
Wednesday, May 8th, 11:30am - 12:45am.
Sala de Cowork, Departamento de Ingeniería Industrial, 2do piso edificio norte. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: Design and Dimensioning of Hydrogen Transmission Pipelines.
Abstract: The global focus on reducing air pollution and dependence on fossil fuels has led to efforts to shift to renewable energy sources. Hydrogen is a promising alternative due to its high energy capacity and ability to regulate electricity production through electrolysis. In this context, the problem of designing and sizing of hydrogen transmission pipelines is presented. The goal is to establish the network topology and diameter dimensions of each pipeline section for hydrogen distribution, to cover the demand at a minimum cost. To address the proposed problem, we consider the dimensioning as the selection of a diameter from a set of available measures, i.e., a discrete diameter approach, and we compare it with a continuous diameter approach from the literature. In our discrete diameter approach, we propose a non-convex quadratic (MIQP) model, and we derive a mixed-integer quadratic convex relaxation (MIQCP). Finally, we adapt a Delta Change heuristic to this context.
Cristian Durán (IMT Atlantique - Nantes)
Phd - LS2N - IMT Atlantique- Nantes, France. Cristian Durán (IMT Atlantique - Nantes)
Friday, April 26th, 11:30am - 12:45am.
Sala de Cowork, Departamento de Ingeniería Industrial, 2do piso edificio norte. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: Exact solution methods for large-scale discrete p-facility location problems.
Abstract: This talk focuses on the exact solution of the NP-hard problems p-median and p-center. These fundamental discrete location problems involve opening a defined number p of facilities and then allocating a set of clients to them according to an objective function to be minimized. Firstly, we study the p-median problem, which seeks to minimize the sum of distances between clients and the open facilities to which they are allocated. We develop an algorithm based on Benders decomposition that outperforms state-of-the-art exact methods by solving instances with up to 238,025 clients and sites. Secondly, we explore the p-center problem, which seeks to minimize the largest distance between a client and its nearest open facility. We study the Benders decomposition and also propose an exact algorithm based on a client clustering procedure based on the structure of the problem. We are able to solve instances with up to 744,710 clients and sites. Finally, we study a robust two-stage p-center problem with uncertainty on node demands and distances. We introduce the robust reformulation of the problem based on the five main deterministic MILP formulations in the literature. We prove that only a finite subset of scenarios from the infinite uncertainty set can be considered without losing optimality. We also propose a column-and-constraint generation algorithm and a branch-and-cut algorithm to solve this problem efficiently.
Luis Rojo (INRIA-Lille)
Phd(c) - INRIA - Lille, France. Luis Rojo (INRIA-Lille)
Wednesday, April 17th, 11:30am - 13:00am.
Auditorium, Industrial Engineering Department, USACH
Title: Electric vehicles and hierarchical optimization models.
Abstract: Recent years have underscored the urgent need for actions to combat global warming, prompting governments to confront this challenge head-on. The transportation sector, as a significant source of emissions, presents a prime opportunity for impactful interventions. This is especially relevant with the rise of electric vehicles (EVs), which necessitate a managed energy supply to ensure the electrical grid operates efficiently. This necessity has given rise to the concept of the coordinated charging problem, which focuses on optimizing EV charging to maintain grid stability. A pivotal aspect of addressing this issue is determining the optimal placement of charging stations, known as the charging facility location problem. These challenges intertwine, creating a complex, chicken-or-egg dilemma that spans both strategic and tactical planning paradigms. Furthermore, the dynamic between various stakeholders, each pursuing their interests in a structured, sequential manner, adds another layer of complexity. This paper delves into the evolution of both the coordinated charging problem and the charging facility location problem, employing hierarchical optimization models to capture the intricate interplay among stakeholders.
Sourour Elloumi (ENSTA-Paris)
Full Professor at ENSTA - Paris, France. (personal website)Sourour Elloumi (ENSTA-Paris)
Wednesday, April 10 th, 11:30am - 13:00am.
Auditorium, Industrial Engineering Department, USACH
Title: On the exact resolution of quadratic programs with integer variables and extensions.
Abstract: A quadratic program in binary variables has very special properties. It can be linearized, i.e. reformulated into a linear program in binary variables, by increasing the number of variables. It can also be convexified, i.e. reformulated into a convex quadratic program in binary variables, by a simple calculation of an extreme eigenvalue. Whether linearized or convexified, the reformulated problem can then be solved by a Branch-and-Bound based on continuous relaxation. These two paradigms have been known for several decades, but neither of them can efficiently achieve the optimum, outside of small or low-density instances. We show that Quadratic Convex Reformulation allows the two paradigms to be put together to strengthen both of them and allow the exact resolution of much larger instances. We then extend this approach to the case where the variables are integers or reals in an interval. We show that it can also be applied to polynomial optimization in binary variables, after a quadratization phase.
Past talks 2023
Past talks 2023
Guido Lagos (UAI - Viña)
Profesor Asistente - Universidad Adolfo Ibáñez - Viña del mar.Guido Lagos (UAI - Viña)
Wednesday, December 27th, 9:00am - 10:00am.
Auditorium, Industrial Engineering Department, USACH
Title: Confiabilidad de sistemas coherentes con fallas simultáneas.
Abstract: En esta charla veremos trabajos recientes de teoría de confiabilidad para sistemas grandes con varias componentes que pueden estar falladas o funcionando, y en que estas fallas pueden producir una falla del sistema. Partiremos revisando algunos conceptos básicos de la teoría de confiabilidad de sistemas coherentes. Estos son sistemas en que más componentes funcionando implican una probabilidad mayor de que el sistema esté funcionando. Luego nos enfocaremos en el modelo de fallas Lévy-frailty Marshall-Olkin, que es una manera de modelar que las componentes puedan fallar simultáneamente. En este contexto, veremos trabajos teóricos recientes sobre aproximaciones de tiempos de falla cuando el sistema es muy grande. También veremos aproximaciones cuando hay costos de reparación y tenemos distintas políticas de reparación.
Nelson Maculan (UFRJ)
Titular Professor - Universidad Federal de Rio de Janeiro - Brasil.Nelson Maculan (UFRJ)
Wednesday, December 6th, 11:30pm - 12:30pm.
Auditorium, Industrial Engineering Department, USACH
Title: Investigación Operativa y la Toma de Decisiones – una Experiencia.
Abstract: El profesor Nelson Manculan, desde 1969, ha estudiado los problemas de decisiones en distintos sectores de las actividades humanas. Tales estudios, lo llevaron a trabajar con modelos matemáticos que buscan soluciones óptimas o que mejoren de cierta forma algún o algunos criterios (objetivos). Es evidente que, para solucionar dichos modelos de optimización, es necesario el uso de los algoritmos de programación matemática (tales como programación lineal, entera, no lineal, grafos, etc.). La especialización académica del profesor Maculan es la programación matemática, junto con la simulación discreta. En sus 56 años de trayectoria, él ha podido participar en muchos proyectos aplicados y en esta ocasión pretende compartirlos con la comunidad.
Fernando Tohme (UNS)
Titular Professor - Univ. Nacional del Sur - Bahía Blanca Argentina.Fernando Tohme (UNS)
Tuesday, December 5th, 15:00pm - 16:00pm.
Auditorium, Industrial Engineering Department, USACH
Title: Industry 4.0: Exploring its Perspectives using Category Theory.
Abstract: The fourth industrial revolution or “Industry 4.0” is a conception of industrial systems that use information extracted in real-time from shop floors and supply chains, gathered through wireless sensors, RFID technologies, Cyber-Physical Systems, and IoT contributing to improving the decision-making processes. Cyber-physical systems (CPS) are the main components of Industry 4.0 production systems. Their interconnection is crucial to the success of distributed and autonomous production plans. The design of global systems, avoiding unwanted results, poses a mathematical problem that we address here in a Category Theory setting. We define a Hypergraph Category of CPS and define a functor representing their behaviors. This ensures that interconnections are accompanied by the composition of the behavior of the connected components. Since this functor is lax we can ensure that the behaviors that may arise will be equivalent to the desired ones. Finally, we apply this framework to formalize Smart Scheduling, showing that the lax functor captures the assignment of operations to cyber-physical systems.
Bibiana Granda (UCM)
Phd(c) - Univ. Nacional del Sur - Universidad Complutense de Madrid.Bibiana Granda (UCM)
Tuesday, December 5th, 16:00pm - 17:00pm.
Auditorium, Industrial Engineering Department, USACH
Title: Transferencia de conocimiento entre universidad e industria. Aplicaciones de la Investigación de Operaciones.
Abstract: Desde la Universidad Complutense de Madrid se desarrollan cierto número de proyectos de consultoría con empresas, ligados a la transferencia de conocimiento entre universidad e industria. Esta charla pretende describir algunos de ellos, mostrando cómo la universidad es partícipe de proyectos con impacto.
Juan Valencia (UAI - USACH)
Research Assistant - UAIJuan Valencia (UAI - USACH)
Wednesday, November 22nd, 11:30pm - 12:30pm.
Auditorium, Industrial Engineering Department, USACH
Title: Is M.L. the best way to manage uncertainty for optimization?
Abstract: Actualmente, muchas de las decisiones en las distintas industrias contemplan incertidumbre, haciendo la toma de decisiones una tarea complicada. Generalmente el pronóstico y la toma de decisiones se modelan como dos pasos secuenciales sin retroalimentación, siendo utilizados de esa forma en áreas como energía, retail, banca, entre otras. Sin embargo, la optimización conjunta de estos procesos puede conducir a mejores decisiones y generar costos más bajos que cuando se consideran de manera separada. En el presente trabajo, presentamos algoritmos para resolver una nueva metodología de aprendizaje en el que los procesos de predicción y toma de decisiones se fusionan, co-optimizado a través de un modelo de optimización de dos niveles para distintos problemas de la literatura, utilizando algoritmos de descomposición los cuales disminuyen la dimensionalidad del problema. Finalmente, mostramos los beneficios de esta metodología a través de experimentos computacionales de optimización estocástica de la literatura.
Tomás Reyes (Pontificia Universidad Católica de Chile)
Profesor Asociado del Departamento de Ingeniería Industrial y de Sistemas (DIIS) de la Pontificia Universidad Católica de Chile.Tomás Reyes (Pontificia Universidad Católica de Chile)
Wednesday, November 8th, 11:30pm - 12:30pm.
Auditorium, Industrial Engineering Department, USACH
Title: Does enhancing the vividness in connection with the future self increase savings behavior? A field experiment.
Abstract: Individuals frequently grapple with the challenge of sufficiently saving for retirement, a problem that can significantly impact the quality of life for retirees. Numerous strategies have been devised to mitigate this issue, ranging from traditional methods such as monetary incentives and tax advantages to more innovative approaches aimed at strengthening the individual's connection with their future self. The latter, though theoretically promising, has not yet been field-tested. The underlying premise is that by amplifying the perceptual vividness of one's future self, individuals might be more inclined to make decisions in line with their long-term interests. This study evaluates this hypothesis through a field experiment involving 415 customers of an investment firm. Participants were randomly assigned into three groups: one without any future self-reference (the control group), a second group presented with a text referencing their future selves, and a third group that was given the same text along with a digitally-aged image of themselves. The results indicate that interventions cultivating a more vivid connection to their future selves increase individuals' intentions to save for retirement. This effect on intentions, however, only translated into a short-term, modest impact on the actual amount of money invested.
Salvador Vicencio (Tecnológico de Monterrey)
PhD(c), Escuela de Ingeniería y Ciencias - Tecnológico de Monterrey.Salvador Vicencio (Tecnológico de Monterrey)
Tuesday, October 24th, 15:30pm - 16:30pm.
Auditorium, Industrial Engineering Department, USACH
Title: A brief tour of the solving methods for the Site-Specific Management Zone problem.
Abstract: Agriculture is essential for the world since it supplies economic and food resources in many countries. In the early 1980s, a Precision Agriculture (AP) concept emerged to highlight the combination of technology and agriculture. Different studies related to AP involve various topics, such as sampling techniques and methods, crop quality, sensor systems, etc. In this talk, we focus on a problem known as the Site-Specific Management Zone problem (SSMZ), which consists of the correct input application in fields with soil variability. To apply the inputs correctly the field should be divided into zones where each zone must be homogeneous according to soil properties that can be chemical or physical. The objective is to minimize the number of management zones. We discuss the methods presented in the literature to solve the SSMZ problem with its advantages and disadvantages. On the other hand, we show a metaheuristic method that improves the results and the computational times existing in the literature. Finally, we introduce a new mixed-integer linear formulation that provides solutions in low computational times for small instances of the problem.
Denise Laroze (Universidad de Santiago)
Profesora Asociada del Departamento de Administración de la Universidad de SantiagoDenise Laroze (Universidad de Santiago)
Wednesday, October 4th, 11:30am-12:30pm
Auditorium, Industrial Engineering Department, USACH
Title: Improving Pension Information: Experimental Evidence on Learning using Online Resources
Abstract: When planning for retirement, deciding what to do with one's pension funds is a high-stakes, one-shot decision, often written in technical jargon that few people understand. It is therefore not surprising that workers experience an important level of stress during the whole pension process, and many of them even prefer to pay for advice or miss out on social benefits they are eligible for because they do not understand the retirement process. We argue that the learning process can be eased by providing information in video format (vs. the standard text) and by changes to the user interface of the websites on which individuals learn about their pension options. The results of both a field experiment with 50 to 70-year-old participants and a lab experiment on university students suggest that videos are significant and substantively more effective in increasing the number of correct responses to a retirement comprehension test. The significance of this effect is robust to changes in the content of the videos, with equivalent impact in private and public pension schemes. The study is conducted in association with the Chilean Superintendencia de Pensiones and experts from the Instituto de Previsión Social.
Víctor Bucarey (Universidad de O'Higgins)
Profesor Asistente en el Instituto de Ciencias de la Ingeniería de la Universidad de O'HigginsVíctor Bucarey (Universidad de O'Higgins)
Wednesday, September 27th, 11:30am-12:30pm
Auditorium, Industrial Engineering Department, USACH
Title: Decision-focused learning through the lens of bilevel optimization
Abstract: Decision-focused learning (DFL) is an emerging paradigm in machine learning which trains a model to optimize decisions, integrating prediction and optimization in an end-to-end system. This paradigm promises to revolutionize decision-making in many real-world applications that operate under uncertainty, where the estimation of unknown parameters within these decision models often becomes a substantial roadblock. This talk discusses bilevel optimization reformulations of the regret-minimization problem that arises under this paradigm. We further show some reformulations and methods to tackle this bilevel optimization problem with off-the-shelve solvers. Finally, we describe the pros and cons of the methods and models studied through a computational study.
Serguey Maximov (University College London)
Investigador Asociado en el Institute for Sustainable Resources, University College LondonSerguey Maximov (University College London)
Wednesday, September 20th, 11:30am-12:30pm
Auditorium, Industrial Engineering Department, USACH
Title: Influencia de las políticas en el financiamiento de energías limpias
Abstract: El proyecto WIND (Wind Investment Network for effective market Design) es parte de una colaboración de UKERC (UK Energy Research Centre) entre UCL, la Universidad de Strathclyde y el Imperial College de Londres, con el objetivo de entender qué políticas de apoyo a la inversión son más eficientes en movilizar las grandes cantidades de capital necesarias para alcanzar la meta del Reino Unido de 50 GW de generación eólica fuera de costa al 2030. Actualmente, los inversionistas enfrentan riesgos significativos relacionados con la "canibalización" de ingresos en mercados marginalistas, debido a la reducción de precios cuando la generación renovable total se acerca o excede a la demanda, lo que disminuye el interés de invertir en el sector. El proyecto busca comprender la influencia de las políticas y el diseño del mercado sobre el riesgo de ingresos, y cómo éste afectan las decisiones de distintos tipos de inversionistas en tecnología eólica fuera de costa.
Yasmín Ríos-Solís (Tecnológico de Monterrey)
Phd | Industrial Engineering - Tecnológico de MonterreyYasmín Ríos-Solís (Tecnológico de Monterrey)
Wed August 30th, 15:00-16:00
Auditorium, Industrial Engineering Department, USACH
Title: Simulación + optimización: minimizar los amontonamientos de buses en el transporte urbano
Abstract: Esta charla se enfoca en mejorar la dinámica de los sistemas de transporte en tiempo real, abordando un problema de amontonamiento de autobuses que se produce cuando los autobuses llegan simultáneamente a una estación. El amontonamiento de autobuses aumenta los tiempos de espera, los tiempos de viaje y genera tasas de ocupación desequilibradas. Hemos implementado una estrategia de simulación conjunta con optimización que implica retener los autobuses por unos minutos en ciertas estaciones una vez que se ha completado el proceso de subida y bajada de usuarios.
Riuwen Lao (Université Paris-Saclay)
Phd candidate | Université Paris-SaclayRiuwen Lao (Université Paris-Saclay)
Wed July 26th, 11:30-12:30
Auditorium, Industrial Engineering Department, USACH
Title: Integrated lot-sizing and energy supply planning with onsite generation of intermittent renewable energy
Abstract: This work considers an industrial production site partially powered by a decentralized energy system based on intermittent renewable energy sources. Our objective is to simultaneously plan the industrial production and the energy supply in this site so as to minimize the total cost. A new way of modelling this combinatorial optimization problem is proposed: it relies on the extension of a multi-product single-resource small-bucket lot-sizing model called the proportional lot-sizing and scheduling problem. This extension involves among others sequence-dependent changeover times overlapping multiple periods and energy-related constraints. Our numerical results show that the resulting mixed-integer linear programming model enables us to obtain good-quality production and energy supply plans with a computational effort much smaller than the one required by a previously published large-bucket lot-sizing model.
Francisco Jara -Moroni (Universidad Diego Portales)
Professor | Escuela de Ingeniería Industrial de la Universidad Diego Portales, Chile | webpageFrancisco Jara -Moroni (Universidad Diego Portales)
Wed July 5th, 9:30-10:30
Auditorium, Industrial Engineering Department, USACH
Title: Métodos matemáticos para problemas de optimización bajo incertidumbre.
Abstract: En esta charla presentaré dos trabajos recientes. El primero, relacionado con mi investigación principal, consiste en el desarrollo de un método para encontrar buenas soluciones locales en problemas de dos niveles estocásticos lineales con solución degenerada en el segundo nivel. Para esto se combinan la Aproximación de Media Muestral (SAA) junto a dos enfoques, uno de busqueda global en una submuestra del conjunto de observaciones, y el otro de penalizaciones duales en el segundo nivel para el problema aproximado. El segundo trabajo se enfoca en encontrar buenas políticas de gestión de energía en residencias con generación via paneles fotovoltaicos y almacenamiento de energía. El trabajo genera distribuciones para radiación solar y demanda energética, mediante técnicas de aprendizaje de máquinas, para luego utilizarlas en un modelo multietapa estocástico que se resuelve mediante la técnica de SDDP. Para ambos trabajos se presentan resultados numéricos que reflejan los beneficios de ambos enfoques.
Felipe Ulloa (University of California)
Ph.d (c) University of CaliforniaFelipe Ulloa (University of California)
Wed May 31st, 11:30-12:30
Auditorium, Industrial Engineering Department, USACH
El presente trabajo, muestra cómo es posible estimar la demanda de transportes de la logística nacional de Chile, medida en toneladas y generadas durante un año, usando un modelo macroeconómico de equilibrio estático como lo es el Modelo de Insumo Producto, propuesto por el premio nobel de economía 1973 Wassily Leóntief.
La modelación que se presenta atiende a las definiciones básicas acerca de qué se entiende por Modelo de Insumo Producto, cuáles son sus componentes, cómo se relaciona con el PIB y cómo es posible, entonces, estimar la demanda de transportes desde de estas matrices.
Posteriormente, se presenta la metodología a emplear, donde se indican los vectores de precio de tonelada promedio que se utilizaran para realizar las transformaciones de las matrices de Leóntief, medidas en flujos de divisas, a matrices medidas en toneladas y como se intercambian esas toneladas entre los distintos sectores económicos.
Se analizan los resultados obtenidos, fijando como base del análisis el uso de la matriz de insumo producto más reciente publicada para Chile por el Banco Central. La Matriz de Insumo Producto 2017, será entonces la base del análisis que se proyecta al 2030 mediante distintas técnicas econométricas, usando la matriz de coeficientes técnicos.
Héctor López-Ospina (Universidad de los Andes)
Professor | Facultad de Ingeniería y Ciencias Aplicadas de la Universidad de los Andes, ChileHéctor López-Ospina (Universidad de los Andes)
Wed May 24th, 11:30-12:30
Auditorium, Industrial Engineering Department, USACH
Title: Incorporación y análisis de la incertidumbre en los sistemas de transporte. Un enfoque basado en modelos de máxima entropía y la lógica difusa
Abstract: Esta investigación extiende la formulación y análisis de modelos de maximización de entropía aplicados a contextos de sistemas de transporte de carga y transporte urbano de pasajeros, incorporando la incertidumbre y ambigüedad de los parámetros.
En sistemas de transporte (pasajeros o carga), la máxima entropía busca la distribución más probable de usuarios (vehículos, personas, carga, productos) con información (parámetros) como demanda, oferta, uso del suelo, flujo vehicular, costos generales de transporte por tipo de usuario, estructura de la red, entre otros. En contextos reales, los parámetros descritos anteriormente pueden ser inciertos, variables o ambiguos. Tal incertidumbre generalmente no se captura ni analiza en detalle en los modelos clásicos. Se utilizan técnicas de optimización difusa para modelar la incertidumbre porque permite resolver adecuadamente problemas de toma de decisiones que involucran múltiples agentes y criterios en entornos inciertos o bajo imprecisión. Además, permite trabajar con información de diferentes fuentes, con subjetividad, integrada con el conocimiento de expertos e información histórica e información faltante (incompleta). Cabe mencionar que capturar la incertidumbre en los procesos de planificación permite comprender y predecir diferentes escenarios y reduce la brecha entre el modelado y la realidad (pero no la elimina). Una de las ventajas de la optimización difusa es que permite trabajar con modelos flexibles, que en escenarios deterministas pueden no tener una región factible dada la complejidad de las restricciones, permitiendo que las condiciones del sistema se cumplan en un determinado nivel. Se presentarán ejemplos numéricos para mostrar las ventajas de los modelos. La presentación terminará con potenciales trabajos futuros en temas de emisiones y externalidades negativas.
Pedro Jara-Moroni (USACH)
Associate Professor | Department of Economics (FAE), University of Santiago. | webpagePedro Jara-Moroni (USACH)
Wed May 10th, 11:30-12:30
Auditorium, Industrial Engineering Department, USACH
Title: Equilibrium Selection in Entry (Global) Games
Abstract: In this work, using a global game approach, we provide conditions to obtain a unique selection of equilibrium in an entry game followed by oligopolistic competition. In such an equilibrium, there is efficient entry, i.e. firms that enter are the ones with the lowest entry cost, providing theoretical foundation for the equilibrium selection assumption utilized in entry models in the empirical entry literature. In particular, we posit a model in which a large number of firms first choose simultaneously whether to enter a market or not, and then firms that decide to enter the market face oligopolistic competition under a parametrized demand. We provide conditions over the primitives of the market, in two classical cases: Cournot competition and price competition with differentiated products. In the talk I will discuss further questions to be addressed.
Nicolás Valenzuela Levi (UTFSM)
Assistant Professor | Departamento de Arquitectura, Universidad Técnica Federico Santa María, Chile.Nicolás Valenzuela Levi (UTFSM)
Wed April 26th, 11:30-12:30
Auditorium, Industrial Engineering Department, USACH
Title: Investigando la Infraestructura Alimentaria Urbana en el Gran Santiago
Abstract: La presentación abordará la investigación que lidera el Dr. Valenzuela Levi en el marco del Fondecyt Iniciación 11221325, la cual se encuentra al inicio del año número 2 de un total de 3. Se explicará el enfoque dado a la investigación, los datos levantados, resultados preliminares y oportunidades de recibir tesistas. Entre los temas a revisar se encontrarán patrones de localización de ferias y supermercados, así como un registro inédito de ollas comunes y distribución de ayudas gubernamentales en la Región Metropolitana, todo con foco en la relación con desigualdades territoriales.
Fabián Fuentes González (USACH)
Assistant Professor | Industrial Engineering Department, University of Santiago, Chile.Fabián Fuentes González (USACH)
Wed April 12th, 11:30-12:30
Auditorium, Industrial Engineering Department, USACH
Title: Financing complex smart local energy systems
Abstract: Smart local energy systems (SLES) are expected to contribute to meeting net zero carbon emission targets, while also addressing issues of energy decentralisation, democratisation, and digitalisation. There are, however, unresolved questions about finance and investment for SLES. We extend an existing corporate governance and risk management model CLASS to a model called CCLLASS. The CCLLASS model is used to qualitatively explore pathways to investment through securitising SLES future cash flows. From a quantitative perspective, we propose combining securitisation and real-options theory to value such systems, account for risks derived from inexperience, innovation or technological change, and offer an asset-backed security to investors entailing a proper incorporation of risks into the pooled cash flows. Case studies evidence is utilised to qualitatively explore governance, risk management practices suited to investor confidence in securitisation. Several governance and risk management measures already in place can support implementation and operation of securitisation, and strengthen investor confidence. Yet, there is scope for improvement in several areas, such as final market architecture and characterisation of benefits to localities. On the other hand, the quantitative aspects are being handled through the Schwartz and Moon (2001) model, as well as Poisson stochastic modelling that aims at upgrading the above-mentioned model by including the default risk, which can be critical when securitising assets and fulfilling obligations with investors.
Talks 2018-2019
Talks 2018-2019
Gonzalo Muñoz (Universidad de O'Higgins)
Assistant Professor | Engineering Institute, Universidad de O'Higgins | webpageGonzalo Muñoz (Universidad de O'Higgins)
Wed Oct 16th, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: Generalized Surrogate Duality for Mixed-Integer Nonlinear Programs
Abstract: The most important ingredient for solving mixed-integer nonlinear programs (MINLPs) to global optimality is a tight, computationally tractable relaxation, which is usually required to be convex. Nonetheless, current optimization solvers can often handle a moderate presence of non-convexities efficiently, which opens the door for the use of potentially tighter non-convex relaxations. In this talk, we show a family of such relaxations obtained via an aggregation of nonlinear constraints. The resulting relaxation is called surrogate relaxation, which is similar to the well-known Lagrangian relaxation but provides stronger dual bounds. We show the benefits and challenges such relaxation can have in general MINLPs and present a generalization of the surrogate relaxation that allows for multiple aggregations of non-convex constraints. We also present an algorithm that can find the best generalized surrogate relaxation, along with several computational enhancements for improving its practical performance.
Renaud Chicoisne (HEC Montréal, Canada)
Postdoctoral Fellow | Department of Logistics and Operations Management, HEC Montréal, Canada | webpageRenaud Chicoisne (HEC Montréal, Canada)
Thu Jul 4th, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: Set-based Lagrangean decomposition methods for mathematical programming
Abstract: We present generic Lagrangean frameworks for primal (variable) and dual (constraint) decomposition algorithms for nonlinear mathematical programs with generalized inequalities. Akin to the Dantzig-Wolfe (DW) method and the Benders Decomposition (BD), we solve a succession of restricted problems/Lagrangean relaxations in a primal setting or relaxed problems/second stage problems in a dual standpoint.
Our approach is generic in the sense that it takes as user-defined inputs
1) a structured subset of the primal (dual) feasibility set, and
2) a mechanism to update it at each iteration.
Depending on the type of subset, the master problems (primal restricted or dual relaxed problem) can have very different structures and properties. We discuss how the structures of the set and the update mechanism influence the number of iterations required to obtain an optimal or near-optimal solution and the difficulty to solve the master problems.
Our meta-algorithm presents a new generic way to prove convergence of - among others - the special cases of DW, BD and recent works arising from an 2009 article of Bienstock and Zuckerberg. We also present linearized schemes to alleviate the computational burden of solving the Lagrangean relaxation in a primal setting or the relaxed problem in a dual setting.
Iván Derpich (USACH)
Professor | Industrial Engineering Department, University of Santiago, ChileIván Derpich (USACH)
Thu Jun 27th, 12:00-13:00
Here's the video of Iván's talk
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: Using Lagrangian Multipliers to Minimize the Contamination in a Warehouse
Abstract: This work addresses the optimal design of warehouses in relation to the contamination for transportation and movement of products and raw materials inside a facility. Two cases are reviewed, the first one when the allocated space is boundless and the second one when the dimensions of the assigned space are given a priori. In both cases a rectangular area is assumed and formulas are designed to calculate the optimal number of storage spaces in the three dimensions, in order to minimize the total energy consumption of transportation over an annual horizon. The developed formulas use a Lagrange function for integrating into a single expression the contamination spending and the constraints associated with the ABC categories and the storage requirements. The developed formulas also deliver the Lagrange multipliers of each constraint, which provides information on the shadow prices of each resource and allow prioritizing investments with respect to the different spaces assigned to each level of the warehouse layout.
Javiera Barrera (Universidad Adolfo Ibáñez)
Associate Professor | Faculty of Engineering, Universidad Adolfo Ibáñez, Chile | webpageJaviera Barrera (Universidad Adolfo Ibáñez)
Thu Jun 20th, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: A decomposition algorithm for computing income taxes with pass-through entities and its application to the Chilean case
Abstract: Income tax systems with “pass-through” entities transfer a firm’s incomes to the sharehold- ers, which are taxed individually. In 2014, a Chilean tax reform introduced this type of entity and changed to an accrual basis that distributes incomes (but not losses) to shareholders. A crucial step for the Chilean taxation authority is to compute the final income of each individual, given the complex network of corporations and companies, usually including cycles between them. In this paper, we show the mathematical conceptualization and the solution to the prob- lem, proving that there is only one way to distribute incomes to taxpayers. Using the theory of absorbing Markov chains, we define a mathematical model for computing the taxable incomes of each taxpayer, and we propose a decomposition algorithm for this problem. This allows us to compute the solution accurately and with the efficient use of computational resources. Finally, we present some characteristics of the Chilean taxpayers’ network and computational results of the algorithm using this network. Join work with Eduardo Moreno and Sebastian Varas.
Andrea Espinoza (USACH)
Assistant Professor | Industrial Engineering Department, University of Santiago, ChileAndrea Espinoza (USACH)
Thu Jun 13th, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: Sustainability: How to implement it?
Abstract: The concept of sustainability refers to meet current needs, without compromising the capacity of future generations to meet their needs, ensuring the balance between economic growth, environmental care and social welfare.
However, what does sustainability really mean? How does it translate into concrete facts? How are these three pillars translated? Currently, there are different tools to measure it, among which are sustainability guides, certifications, criteria and indicators, standards and policies. Unfortunately, we can observe an ex-post evaluation approach mainly, i.e., these tools are focused on the measurement of sustainability of implemented companies and/or productive processes.
Therefore, it is necessary to make an abstraction of these tools in order to design sustainable enterprises and/or production processes. Which can be done from the industrial engineering, through the multiobjective optimization.
Jorge Zamorano (USACH)
Assistant Professor | Industrial Engineering Department, University of Santiago, ChileJorge Zamorano (USACH)
Here's the video of Jorge's talk!
Thu May 23rd, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: Environmental regulation under imperfect competition
Abstract: In this talk I will give an overview of two of my research tracks in the field of regulation of economic activities that harm the environment: one is the design of pollution permits and the other is waste management.
Regarding the design of pollution permits, I will talk about two papers. One paper analyses the implementation of pollution permits. It focuses on the distributional impacts linked with the stringency of output-based allocation, when the market for permits covers two sectors and the total cap is held constant. Another paper addresses the issue of differentiating permit allocation across areas, this being linked to the possibility of firms to relocate.
Regarding waste management, I will talk about a working paper that compares the efficiency of extended producer responsibility (EPR) programs and the efficiency of an ex-ante tax. The tax allows more ex-ante flexibility regarding market conditions, but the EPR allows more ex-post adaption to cost realizations. As a result, the relative efficiency of the EPR increases with uncertainty of the costs and competitiveness of the market. We propose several extensions of this basic result. Also, we propose a new partial EPR solution and delineate the circumstances under which this new policy instrument can outperform the traditional ones.
Diego Morán (Universidad Adolfo Ibáñez)
Assistant Professor | Business School, Universidad Adolfo Ibáñez, Chile | webpageDiego Morán (Universidad Adolfo Ibáñez)
Here's the video of Diego's talk!
Thu May 16th, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: Some properties of convex mixed-integer programs
Abstract: Mixed-integer programs (MIPs) are optimization problems where some of the decision variables are required to take integer values. Despite the usefulness of linear MIPs in solving real-world problems, there are many engineering and business applications that cannot be modeled by using linear constraints only. Therefore, there is a need to understand basic properties of MIPs that require more intricate constraints, such as nonlinear convex constraints.
In the first part of the talk, we briefly present some advances in various fundamental concepts in convex MIPs, such as properties of their feasible region and properties of cutting planes. In the second part of the talk, we present in more detail our extension of the subadditive duality theory for linear MIPs to the more general case of conic MIPs. This work is motivated by the algorithmic and theoretical implications that this 'nice' property has in the linear case.
This is joint work with Gustavo Angulo (PUC), Santanu S. Dey (Georgia Tech), Burak Kocuk (Sabanci U.) and Juan Pablo Vielma (MIT).
Luis Rojo & Nelson Troncoso (USACH)
MSc Students, Master in Industrial Engineering Sciences | University of Santiago, ChileLuis Rojo & Nelson Troncoso (USACH)
Thu May 9th, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: The problem of estimating transition rates in the dynamic modeling of body mass index in the Chilean population
Abstract: This work presents a non-linear programming model which represents a dynamic model of obesity prevalence on population of Chile, considering peoples’ trend to modify their Body Mass Index by each belonging age group. The goal is to minimize weighted mean square error among real and estimated population distribution, using data obtained from Encuesta Nacional de Salud 2009/2010. There is a phase of data fitting as population fit according to log-normal distribution and forecast of mortality and birth rates year-by-year. Also, a sensitivity analysis is performed by several times according to uniform distribution between normal probability of occurrence equal to 62.5% and 99.5%. We propose many repetitions modifying Solver seed for problem resolution due to non-linearity of the problem. Obtained results are compared with Encuesta Nacional de Salud 2016/2017 to analyze and to validate them. Population forecasting by estimated trends gives an average error equal to 0.36% and 0.79% for males and females, respectively. Finally, an underlying result show Obesity classification, according to World Health Organization, will be dominant; where 39.29% of males and 48.68% of females will have this nutritional status.
Sebastián Dávila (USACH & Universidad de Chile)
Associate Researcher, University of Santiago, Chile | PhD Candidate, Doctorado en Sistemas de Ingeniería, Universidad de Chile, ChileSebastián Dávila (USACH & Universidad de Chile)
Thu May 2nd, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: Bilevel Programming Model for a Multi-Product Location Problem
Abstract: A retail firm has several malls with a known location. A particular product, e.g., food processor, comes in types, which differ by shapes, brands and features. Each mall has a limited capacity of products to be sold at that location, so the firm has to choose what products to sell at what mall. Furthermore, the firm can apply discrete levels of discount on the products, e.g., 5% and 10% over the price. The objective of the firm is to find what products to sell at which mall, with what level of discount, so that its profit is maximized.
Consumers are located in points of the region. Each consumer or group of consumers has a different set of acceptable products, and will purchase one of these, or none if it is not convenient for her. Consumers maximize their utility.
The agents (firm and consumers) play a von Stackelberg game, in which the firm is the leader and the customers the follower. Once the firm decides the products to sell at each mall and the possible discounts, consumers purchase (or not) one of their acceptable products wherever their utility is maximized.
We model the problem using first a bilevel formulation, and we further replace the follower problem by the primal constraints and optimality restrictions, to obtain a compact formulation. We also present a strong and a weak formulation.
Computational experiments are shown, using known instances from the literature.
Andrea Canales (Universidad de O'Higgins)
Postdoctoral Researcher | Universidad de O'Higgins, ChileAndrea Canales (Universidad de O'Higgins)
Thu Apr 25th, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: Collusion in contestable markets
Abstract: We consider a price competition model in which a few large firms acting as a cartel, share the market with a fringe of many small suppliers in an infinite-horizon setting. In each period the cartel sets the price and small firms, incumbents or future entrants, observe it and decide if stay or out of the market. We study the sustainability of collusive agreements for large firms using trigger strategies. The main result is that, under certain conditions, a punishment price less aggressive than Bertrand price makes the agreement more sustainable, in the sense of it can be sustained by a smaller discount factor. The economic intuition behind this result is that, if a member of the cartel deviates from the collusive agreement, it can get profit from the others member of the cartel and also from the competitive fringe which decide to leave the market anticipating the even lower price in the future.
Felipe Lagos (Georgia Tech)
PhD Student | School of Industrial & Systems Engineering, Georgia Institute of Technology, USA | webpageFelipe Lagos (Georgia Tech)
Thu Jan 3rd, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Sign up here to get a free sandwich+beverage at the talk!
Title: Inventory routing problem in continuous time: a method with optimality guarantees
Abstract: In time dependent models the objective is to find the optimal times (continuous) at which activities occur and resources are utilized. These models arise whenever a schedule of activities needs to be constructed. One common approach consists of discretizing the planning time horizon in a time-expanded network and then limit the decisions to those time points. However, this approach leads to very large formulations that are intractable in practice. In this work we study a particular time dependent model, the Continuous Time Inventory Routing Problem (CIRP). In this problem a company manages the inventory of its customers, resupplying a single product from a single facility during a finite time horizon. The product is consumed by a constant rate (product per unit of time) by each customer. The customers have local storage capacity. The goal is to find the minimum cost delivery plan that ensures that none of the customers run out of product during the planning period. In this work, we propose a novel method to solve the CIRP based on partially constructed time-expanded networks. This method iteratively discovers time points needed in the network to find optimal solutions. We test this method using random generated instances. This is the first time that provable optimal solutions are presented for the CIRP.
This is joint work with Martin Savelsbergh and Natashia Boland.
Daniela Hurtado (Georgia Tech)
PhD Student | School of Industrial & Systems Engineering, Georgia Institute of Technology, USA | webpageDaniela Hurtado (Georgia Tech)
Thu Dec 20th, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: Transform Methods for Heavy-Traffic Analysis
Abstract: Heavy-traffic analysis of queueing systems has been done for years using the diffusion limits approach. More recently, the drift method was as an alternate approach. In this method, drift of certain test functions is set to zero in steady-state to obtain the steady-state moments of the queue lengths. Then, the complete distribution of the heavy-traffic scaled queue lengths is obtained from the moments. In this talk, we present a novel view of this method. We show that the drift method can be thought in terms of the moment generating function or characteristic function of the heavy-traffic scaled queue length in steady-state. This view unifies the drift method with the recently developed BAR approach for heavy-traffic limits. We present this view for systems where the so called complete resource pooling condition (CRP) is satisfied, and reprove known results. Extending this method when the CRP is not satisfied, such as the case of an input queued switch, is work in progress.
You can download Daniela's presentation here
Andrés Fielbaum (Universidad de Chile)
PhD Student | Doctorado en Sistemas de Ingeniería, Universidad de Chile, Chile | webpageAndrés Fielbaum (Universidad de Chile)
Thu Dec 13th, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: Optimal design for a public transport bus system considering routes, frequencies and vehicle sizes in a general city model
Abstract: Several lines compose a transit bus network, each of them characterized by their route, frequency and vehicle size. The problem of finding the optimal set of lines for a given graph and demand pattern is quite complex, and it has given raise to several strategies to tackle it. In this presentation we will give a general introduction to this problem, to then explain how we faced it. For this, we will first describe a simple city model (a graph with the corresponding users flows between its nodes) that has been proved useful for transport analysis and that allows for representing monocentric, polycentric and dispersed cities. Then, we will propose and optimize four strategic line structures over this city model, and compare them with the line structures that emerge when applying four different heuristics. We will identify clearly the relation between the urban setting with the lines structure that shows the best response. Finally, we will analyze how the operational characteristics of the system vary with the number of users, to identify new sources of scale economies induced by the line structures optimization process.
You can download Andres' presentation here
Francisco Jara Moroni (USACH)
Assistant Professor | Industrial Engineering Department, University of Santiago, Chile | webpageFrancisco Jara Moroni (USACH)
Thu Dec 6th, 12:00-13:00
Postgrado B room, 2nd floor North building, Industrial Engineering Department, USACH
Title: Mathematical Programs with Complementarity Constraints: Methods and Applications
Abstract: Mathematical Programs with Complementarity Constraints (MPCCs) are a particular family of optimization problems where among its constraints there is the requirement that two non-negative vectors must be orthogonal (complementary) to each other. This set of constraints is what makes the problem difficult to tackle, since no feasible point satisfies standard constraint qualifications. Due to this, common non-linear programming methods tend to perform quite poorly.
MPCCs arise in several areas including hierarchical programming, Stackelberg games, mechanism design and quantile minimization, among others.
In this talk, I will go through several applications for MPCCs, mostly in the context of bilevel programming and chance constrained problems, and explain in some detail some methods used to solve them to global or local optimality.
Guido Lagos (USACH)
Assistant Professor | Industrial Engineering Department, University of Santiago, Chile | webpageGuido Lagos (USACH)
Wed Nov 28th, 12:00-13:00
Auditorium, Industrial Engineering Department, USACH
Title: Systemic risk & reliability in networks: asymptotic analysis and confidence bounds
Abstract: Systemic risk in networks studies the risk of collapse of the entire system due to failure, shocks or actions of components of the network. This area has received particular attention in the last decade or so due to the need to better understand recent financial and economical crises. In this context, it is of interest to analyze, from a probabilistic perspective, the instant at which the whole system is considered to fail, in the case where there are failures that simultaneously affect several components of the network.
In this work we take an unexplored approach in the field of systemic risk and study the following question: Are there asymptotical regimes for the probabilistic behavior of the network that allow to approximate the system and ultimately say something about the reliability of the network? Here, in broad terms, we refer to "asymptotic regime" as an equilibrium distribution or state of the system as the size of the network and time grow together in a "balanced" way. We answer the previous question in the positive, in the case where the dependency model for the failures is a particular exponential Marshall-Olkin copula and when the whole system is considered to fail when the last component fails — or more generally when the k-th component (out of a total of n components) fails.
Our main contribution is a collection of results giving a detailed analysis of a non-trivial scaling regime for the probability of the network being working at a certain time, as the time and size of the network scales. These approximations allow to estimate and give confidence bounds for the failure probabilities of the system. Moreover, the dependency model we study only needs to specify a reduced number of parameters. All in all, our results allow to study the common-cause failure models on networks in a realistic, relevant, and practical fashion.
Joint work with Javiera Barrera (Universidad Adolfo Ibáñez)
Ángela Flores - Universidad de Chile
Phd - Departamento de Ingeniería Eléctrica - Universidad de Chile.Ángela Flores - Universidad de Chile
Lunes, 18 de Noviembre, 11:30am - 12:45am.
Auditorio, Departamento de Ingeniería Industrial. Av. Víctor Jara 3769, Metro U. de Santiago.
Title: Power system planning under multi-scale climate-induced uncertainty
Abstract: Power systems are facing increasing pressure from climate change. The rise in global temperatures, more variable precipitation patterns, and the increase in frequency and severity of extreme weather events pose substantial risks for power systems across multiple time scales. In this talk, a comprehensive power system planning model is introduced to anticipate both short- and long-term uncertainties by leveraging multi-scale stochastic programming. Furthermore, a novel approach is presented for constructing multi-scale scenario trees to represent climate-induced uncertainty, using historical weather data and future climate projections. Finally, suitable decomposition methods for capacity planning under uncertainty and possible enhancements are explored.
Report abuse