Transportation infrastructure assets deteriorate over time due to natural hazards, heavy traffic, and aging, increasing their risk of failure. National transportation agencies must strategically invest in maintenance to avoid significant social and economic impacts. We address the infrastructure maintenance planning problem, in which a maintenance plan must be designed for each asset within a budget limit to maximize the weighted average asset condition over a planning horizon. We derive a knapsack-type mathematical formulation and propose an exact column enumeration algorithm to solve it. First, a column-and-cut generation algorithm computes a (dual) upper bound on the optimal value. The master problem selects a maintenance plan for each asset and is strengthened with extended q-cover inequalities. By representing maintenance plans as paths over a directed acyclic multigraph that captures asset deterioration and maintenance decisions, the pricing problems unveil feasible plans through a specialized labeling algorithm. Second, a relaxation-enforced neighborhood search finds a (primal) lower bound. Finally, using these bounds, we enumerate sufficient columns to find an optimal solution via a commercial MILP solver. Computational results on generated instances spanning a 10-year planning horizon demonstrate that our algorithm delivers optimal solutions for instances with up to 50 assets and near-optimal solutions (gap < 0.18%) for instances with up to 100 assets within five hours.

This study demonstrates that advanced optimization modeling and algorithm design can directly improve how society invests in and preserves critical infrastructure, ensuring safer and more reliable systems for the future. You can read the full article at the following link: https://www.sciencedirect.com/science/article/abs/pii/S0305054825002758?via%3Dihub

En el marco del evento, se otorgaron distinciones a los trabajos más sobresalientes presentados. Esteban Leiva, estudiante de la Maestría en Ingeniería Industrial y egresado del pregrado en Matemáticas, recibió una mención de honor por la calidad y rigurosidad de su trabajo, el cual fue evaluado junto a contribuciones de estudiantes de doctorado provenientes de diversas instituciones internacionales. Este reconocimiento destaca no solo la excelencia académica del estudiante, sino también el nivel de formación e investigación que se promueve en programas de posgrado de la región.

Para más información, visita: https://optimization-workshop.github.io/

El Workshop de Optimización: Teoría, Algoritmos y Aplicaciones se llevó a cabo en diciembre de 2024 en Bogotá, Colombia, en la Universidad de los Andes. El evento reunió a un grupo diverso de investigadores y estudiantes de instituciones locales e internacionales, con el objetivo de fortalecer la colaboración en el campo de la optimización matemática y promover la investigación en la región, especialmente entre jóvenes investigadores.

Uno de los conferencistas principales fue el profesor Andrés Medaglia, director del grupo de investigación COPA, quien presentó la charla titulada “Aprovechando Estructuras de Caminos Más Cortos para Soluciones de Ingeniería”.

El workshop fue ampliamente reconocido como un éxito, ofreciendo una plataforma dinámica para el intercambio de ideas y el desarrollo de nuevas colaboraciones en optimización. Este evento representa un aporte significativo para la consolidación de una comunidad de investigación más sólida en este campo, tanto en Colombia como en la región.

Para más información, visita: https://optimization-workshop.github.io/

COPA participated in the 2024 INFORMS/ALIO/ASOCIO International Conference, held from June 16th to 19th in Medellín, Colombia. This leading international event brings together top academics and practitioners in Operations Research and analytics from around the world to share cutting-edge research and developments.

The following (current and past) COPA members presented at the conference, listed in alphabetical order along with the titles of their presentations:

• Alfaima L. Solano-Blanco: "Sustainable crop planning for mixed cropping systems – An optimization approach"
• Andrés Gómez: "Robust support vector machines via conic optimization"
• Andrés L. Medaglia (Plenary): "A (Not So) Shortest Path: Models, Solutions, and Applications"
• Daniel Yamín: "Branch-Price-and-Cut for Infrastructure Maintenance Planning"
• Juan G. Villegas: "A Bilevel Vehicle Routing Problem for a Colombian Package Delivery Company"
• Julio Goez: "Assessing the Effect of Consolidation of Freight Transport: A Case Study in Norway"
• Luis Zuluaga: "Constraint Penalization for Non-convex Optimization Problems: Applications to Quantum Computing"
• Natalia Pacheco-Carvajal: "Automatic Classification of Academic Articles via Machine Learning"

You can find some pictures with their corresponding description below:

1. Top Left: Natalia Pacheco-Carvajal
2. Top Right: Michel Gendreau and Andrés L. Medaglia
3. Bottom Left: Andrés L. Medaglia and Juan G. Villegas
4. Bottom Center: Daniel Yamín
5. Bottom Right: Alfaima L. Solano-Blanco

For more information, visit https://meetings.informs.org/wordpress/2024international/ or contact us via email at Esta dirección de correo electrónico está siendo protegida contra los robots de spam. Necesita tener JavaScript habilitado para poder verlo..

School infrastructure is crucial for providing quality education. However, natural disasters like earthquakes, hurricanes, floods, and landslides can damage school facilities. In addition, developing countries often face infrastructure challenges like inadequate classrooms, poor lighting, and insufficient ventilation. To tackle these challenges, COPA and CIMOC researchers together with researchers from UCL’s UNESCO Chair in Disaster Risk Reduction and Resilience Engineering have recently published three papers on this topic to address these issues on a large scale.

The first paper, "Towards Disaster Risk Mitigation on Large-Scale School Intervention Programs", proposes a data-driven methodology for prioritizing interventions to reduce seismic disaster risks in schools. The method uses a clustering algorithm to group similar buildings based on a seismic taxonomy and designs scalable interventions applicable to the entire cluster. You can access the paper here: https://doi.org/10.1016/j.ijdrr.2023.103655

The second paper, "A Decision-Making Framework for School Infrastructure Improvement Programs", proposes a decision-making framework for prioritizing school infrastructure investments in safety and functionality with limited budgets. It uses clustering procedures, a multi-criteria utility function, and an optimization component to improve school infrastructure holistically. You can learn more about the framework here: https://doi.org/10.1080/15732479.2023.2199361

The third, and most recent, publication titled "Large-Scale School Building Infrastructure Improvement: The Case of the City of Cali, Colombia" presents an extension of the decision-making framework. This extended framework can be used in the early stages of formulating school-building improvement programs. The paper is implemented as a case study of public school infrastructure in the city of Cali, Colombia. The method helps define the need for improvement and the budget required to implement a large-scale intervention program for different decision-maker profiles. The paper can be accessed via the following link: https://doi.org/10.1016/j.seps.2024.101881

According to the authors, this research is significantly relevant because it provides robust methodologies to support decision-making for school infrastructure improvement. This approach optimizes investments, improves infrastructure quality, adds transparency, and can reduce corruption by standardizing the allocation process. Additionally, the proposed methods are easily adaptable to different countries, levels of information quality and quantity, and even other types of infrastructure, making it ideal to be applicable in different contexts.

Diferentes generaciones de COPA se encontraron en un seminario en Montréal. En la foto, de izquierda a derecha, está Jorge E. Mendoza (profesor asociado en HEC Montréal), Leonardo Lozano (profesor U. of Cincinnati), Nicolás Cabrera (estudiante doctoral en HEC Montréal) y Ariel Rojas (pasante en HEC Montréal). Mucho talento junto en esa foto con investigadores COPeros que, entre muchas otras cosas, han creado y extendido las ideas del Pulse Algorithm (ver: https://ifors.org/newsletter/ifors-news-june-2018.pdf).