EPF webinars are organized monthly and free, follow the EPF Eventbrite page to be notified when booking is live and get updates on the upcoming webinars.
Previous webinars:
EPF webinars are organized monthly and free, follow the EPF Eventbrite page to be notified when booking is live and get updates on the upcoming webinars.
Previous webinars:
Speaker: Ryan O’Neil, CTO and Co-Founder of Nextmv
More and more, data science and decision science practitioners are seeking to combine machine learning forecasts with actionable and optimized decisions. This can include anything from predicting traffic patterns for delivery scheduling to consumer buying behavior for inventory management. But bridging these two disciplines can be challenging.
In the on-demand logistics space, these worlds are colliding more frequently with practitioners generating demand forecasts that feed into shift scheduling models that feed into vehicle routing models. Getting to an 80% good solution for the optimization side is not hard. What is hard is the remaining 20% where people tend to over-optimize their models using fixed inputs (which makes the model more brittle in the face of uncertainty). What if, instead, we could take those 80% solutions and use horizontal compute to scale them up in the face of uncertainty?
In this talk, we will explore what has made blending ML and OR outputs challenging, the roles of deterministic and stochastic optimization in relation to ML, and how scenario testing techniques via horizontal computing provide an expedient and more accessible path for combining these worlds.
Ryan O’Neil is CTO and cofounder of Nextmv (nextmv.io). Previously, he led the Decision Engineering department at Grubhub and Zoomer, which owned forecasting, scheduling, routing, and simulation. Ryan worked as an Operations Research Analyst at MITRE, and led software teams at The Washington Post, Yhat, and Polimetrix. During this, he earned a PhD in Operations Research at George Mason University, and wrote his dissertation on real time routing for pickup and delivery problems.
EURO Practitioners’ Forum past and planned activities are available to the Forum members, as well as the wider public.
Visit the website and register as a member for free, to get the regular updates on all activities: EPF Member registration page. The recordings and details from previous webinars are also available on this website.
Follow the Forum on Twitter, LinkedIN and Eventbrite.
EURO Practitioners’ Forum webinars organisers: Sofiane Oussedik (IBM), Joaquim Gromicho (ORTEC), Torkel Haufmann (Sintef), Adisa Mujezinovic
Speaker: Ruth Kaufman, co-founder, Pro Bono OR and EURO Practitioners’ Forum chair
Join us for Doing OR for free: the UK OR Society’s Pro Bono OR service, where you can volunteer your skills to help organizations in need!
Pro Bono OR connects charitable organisations with skilled OR analysts who are willing to give their time for free. Our volunteers use a broad array of OR techniques and approaches to help UK-based NGOs identify their critical issues, explore different options, and explain their impact. We give executives the power to make more effective decisions and build more impactful organisations.
Come along and learn:
We will start with a brief review of what sort of skills OR people offer to non-profit clients, and what sort of organisations we serve. We will use case studies and examples to demonstrate the difference we’ve made for NGOs and volunteers over our 11 year history.
3 MAY 2024 WEBINAR RECORDING
EURO Practitioners’ Forum past and planned activities are available to the Forum members, as well as the wider public.
Visit the website and register as a member for free, to get the regular updates on all activities: EPF Member registration page. The recordings and details from previous webinars are also available on this website.
Follow the Forum on Eventbrite, Twitter and LinkedIN , and feel free to get in touch.
EURO Practitioners’ Forum webinars organisers: Sofiane Oussedik (IBM), Joaquim Gromicho (ORTEC), Torkel Haufmann (Sintef), Adisa Mujezinovic
Speaker: Dr. Antti Korhonen, CEO, Detech Decision Technologies, Finland
The presentation discusses optimization solutions for strategic financial decision-making and enterprise-wide risk management in financial institutions. The software is a sophisticated prescriptive analytics solution based on a stochastic dynamic optimization approach and solves highly complex real-world financial planning problems under uncertainty. The application is an integrated platform for banking, insurance, pension funds, wealth management and multinational financial groups.
The software computes dynamic, multi-period financial strategies with concrete action plans that fulfill diverse management goals, priorities and regulatory requirements. The strategies help institutions to prepare simultaneously for multiple economic and financial market scenarios in uncertain operating environments. The computed forward-looking strategies lead to considerable improvements in performance when compared with more traditional approaches. This improvement in performance becomes particularly noticeable in exceptionally difficult and turbulent operating environments.
5 APRIL 2024 WEBINAR RECORDING
EURO Practitioners’ Forum past and planned activities are available to the Forum members, as well as the wider public.
Visit the website and register as a member for free, to get the regular updates on all activities: EPF Member registration page. The recordings and details from previous webinars are also available on this website.
Follow the Forum on Eventbrite, Twitter and LinkedIN , and feel free to get in touch.
EURO Practitioners’ Forum webinars organisers: Sofiane Oussedik (IBM), Joaquim Gromicho (ORTEC), Torkel Haufmann (Sintef), Adisa Mujezinovic
Speaker: Joaquim Gromicho, Professor of Business Analytics at the University of Amsterdam and Science and Education Officer at ORTEC
Joaquim A.S. Gromicho acts as Science and Education Officer for ORTEC and is full professor of Business Analytics at the University of Amsterdam. He received his PhD in Optimization in 1995 from the Erasmus University Rotterdam, before spending two years as Assistant Professor at the University of Lisbon. He serves the Dutch Statistics and OR Society as editor in chief of STAtOR, a magazine on applications and impact, and the steering committee of the EURO Practitioner’s Forum.
In the era of generative AI and powerful ML packages prevalent in the Python ecosystem, the significance of Mathematical Optimization remains paramount. Recognizing the historical confinement of this tool to the realm of Operations Research experts, Krzysztof Postek, Alessandro Zocca, Jeff Kantor, and I have embarked on a mission to democratize its accessibility. In our collaborative effort, we’ve authored a hands-on, accessible yet ambitious manual, tailored for a diverse audience, including students and professionals. This comprehensive guide not only delves into traditional (mixed-integer) linear optimization but also explores network, convex, conic, stochastic, and robust optimization.
Scheduled for release by Cambridge University Press as an inexpensive textbook in 2024, our book, also available under a Green Open Access license, is designed to empower learners with practical insights. The latest prepublication version is accessible at https://ortec.com/en/featured-insights/insights/math-optimization-python. Realistic examples accompany each topic, and to further enhance the learning experience, we provide an online companion. This resource comprises 50+ fully functional notebooks readily deployable in Google Colab, serving as practical starting points to address diverse real-life challenges you may encounter.
For your convenience, the companion is available at https://mobook.github.io/MO-book/intro.html, within the permissive MIT license. Note that AMPL ported it to their own modelling language at https://mo-book.ampl.com/.
During this webinar, I will demonstrate representative content of the book in action and will also showcase how ChatGPT serves as a valuable collaborator in my professional endeavors, offering insights into its role in my workflow.
2 FEBRUARY 2024 WEBINAR RECORDING
EURO Practitioners’ Forum past and planned activities are available to the Forum members, as well as the wider public.
Visit the website and register as a member for free, to get the regular updates on all activities: EPF Member registration page. The recordings and details from previous webinars are also available on this website.
Follow the Forum on Twitter and LinkedIN , and feel free to get in touch.
Speaker: Carlos Henggeler Antunes, INESC Coimbra, Department of Electrical and Computer Engineering, University of Coimbra, Coimbra, Portugal
Demand response programs play a key role in the energy transition, offering potential benefits to multiple players: grid operators (alleviating congestion in networks), retailers (enabling to manage wholesale buying and retail selling prices) and consumers (reducing the energy bill without jeopardizing comfort). This work presents a modular set of mixed-integer linear programming models of appliance operation aimed at being incorporated in autonomous Home Energy Management Systems allowing for the integrated optimization of all energy resources. The models consider different cost objective function components (energy and power costs, monetized discomfort) as well as dealing with the user’s discomfort derived from operating appliances out of the habitual periods, settings, and temperature ranges. The models are developed in a building block manner to be used with different objectives of economic and comfort nature.
12 JANUARY 2024 WEBINAR RECORDING
EURO Practitioners’ Forum past and planned activities are available to the Forum members, as well as the wider public.
Visit the website and register as a member for free, to get the regular updates on all activities: EPF Member registration page. The recordings and details from previous webinars are also available on this website.
Follow the Forum on Twitter and LinkedIN , and feel free to get in touch.
Speaker: Claude Philippe Medard, Principal Data Scientist, SAP
We discuss solutions to two business processes performed during Airport/Airline Ground Operations.
The first is the challenging task of optimal allocation of Stands and corresponding Tow plans for a large number of aircrafts flying in and out of airports. The second task is the subsequent planning and monitoring of Turnaround processes required for orchestrating all ground handling activities.
The presentation will showcase solution approaches to these highly time-critical operations, both for planning and recovery. The solutions adhere to strict dependency rules as required by operations and provide improved accuracy for Target Off Block Time slot requests to Air Traffic Control (ATC) by incorporating real-time IOT data, ML predictive algorithms for improved duration estimates and Optimization for optimal allocation while balancing multiple objectives.
The study was done with a large commercial airline with data from their productive scenarios.
1 DECEMBER 2023 WEBINAR RECORDING
EURO Practitioners’ Forum past and planned activities are available to the Forum members, as well as the wider public.
Visit the website and register as a member for free, to get the regular updates on all activities: EPF Member registration page. The recordings and details from previous webinars are also available on this website.
Follow the Forum on Twitter and LinkedIN , and feel free to get in touch.
Speaker: Ana Sofia Pereira
The vehicle routing problem (VRP) is widely known in the OR literature. Several variants are well-studied but what about the real-world constraints found when actually deploying such algorithms at transportation companies?
Routyn is a VRP solution provider and throughout this session we’ll share some of the constraints found on companies across the world over the past 20 years of real-world experience.
3 NOVEMBER 2023 WEBINAR RECORDING
EURO Practitioners’ Forum past and planned activities are available to the Forum members, as well as the wider public.
Register as a member for free, to get the regular updates on all activities: EPF Member registration page. The recordings and details from previous webinars are also available on our website
Follow the Forum on Twitter and LinkedIN , and feel free to get in touch.
Speaker: Nicolás García Aramouni, Data Science Associate Manager, Accenture
Operations research and analytics have experienced significant growth in society, but their adoption in developing countries like Argentina has been slower. This is particularly evident in sectors like public schools, where outdated pen-and-paper or spreadsheet methods persist, hindered by low budgets. In Buenos Aires, three schools manually create schedules, a process that can take weeks or months due to multiple factors. However, these schools lack the necessary skills and resources for analytics-based solutions.
This study presents a pro-bono project aimed at helping these schools develop a user-friendly scheduling solution. The project involves constructing a Mixed Integer Programming (MIP) Model that considers teacher availability and subject hours, reducing the schools’ manual work. The model incorporates basic and tailored constraints, addressing the challenge of scheduling teachers across multiple schools. By implementing this solution, the schools achieved record time efficiency in producing schedules for the 2023 year. This project showcases the power of analytics-based solutions and underscores the role of businesses in supporting their communities through pro-bono initiatives.
Abstact
The role of operations research and analytics in our society has been growing in the last few years at an increasing pace. However, in developing countries like Argentina, the adoption of these methodologies has been slower and is currently being led by businesses and universities. In this scenario, some “old-fashioned” sectors still base most of their problem solving on pen-and-paper or spreadsheet software, which makes change management more difficult. This is the case of public schools, a sector which historically has had important infrastructure issues, both from a physical and systems standpoint, generated by low budgets. Under this scenario, teachers usually don’t have high salaries, making them have jobs in multiple schools and partitioning their daily availability in multiple portions that satisfy each institution’s necessities. This generates an interesting challenge when creating the school schedule as teachers do not have exclusivity for the school in question. When talking to three schools in Buenos Aires, Argentina, they told us that currently they do the complete scheduling process manually, which usually takes up multiple weeks or even a whole month at times, which makes sense as they must consider a great number of factors and variables all at once, in addition to completing other tasks, unrelated to this process.
These schools don’t have employees with the skills necessary to help them to create an analytics-based solution that solves this problem. In addition, people with this knowledge are usually skilled workers who usually can get higher salaries working for a technology company, making the situation more difficult for the budget-constrained schools. Therefore, businesses should have a more active role in society, helping their community and sharing their specific knowledge through pro-bono activities.
The idea of this pro-bono project is just that: helping three schools in Buenos Aires create a schedule in a user-friendly way. To achieve this, we created a solution that had two very important objectives:
1) Build a Mixed Integer Programming (MIP) Model that generated a valid schedule considering the availability of each teacher, how many hours a subject needs, etc.
2) Change and reduce the manual work done by schools In this sense, our aim was to change the work made by these schools to create the schedule: instead of having to create the whole schedule, they would only need to manually create the inputs needed by the model. We asked them to create two inputs: one would
indicate how many hours each subject needs to be taught at each course, and who is the teacher of that course. The other one should reflect the availability of each teacher. These inputs were fed into our MIP model, which incorporated both basic constraints (a teacher cannot teach two different subjects simultaneously) and tailored constraints (for example, certain subjects couldn’t be taught when students just got to school at 8 AM). Considering that multiple teachers were employees of the three schools, we couldn’t partition the problem, generating the challenge of building a model that could generate the schedule for all.
Here, there is no revenue metric that we are trying to optimize. However, what we did was improve productivity as this model let these schools produce their schedules for the 2023 year in record time, as they only needed to invest a couple of days to generate these inputs and let us know the conditions and constraints that the final schedule should follow. As an additional result, we showed the power of analytics-based solutions and were able to share this with our local community.
EURO Practitioners’ Forum past and planned activities are available to the Forum members, as well as the wider public.
Visit the website and register as a member for free, to get the regular updates on all activities: EPF Member registration page. The recordings and details from previous webinars are also available on this website.
Follow the Forum on Twitter and LinkedIN , and feel free to get in touch.
Presented by Dr. Inci Yüksel-Ergün, Zuse Institute Berlin (ZIB)
Data is ubiquitous in the age of analytics. The reliability of decisions based on OR studies depends on the underlying data quality. However, identifying pertinent data and assessing its quality is challenging. It is inevitable to employ highly-connected and consistent real-world data sets to model complex decisions. When expert knowledge becomes obsolete with disruptive changes, we require more complex models to comprehend the impacts of these changes.
When conducting projects with industry using highly connected data, we encountered several cases where our analysis detected data errors that were too complex for humans to understand. Examples for our analysis include irreducible infeasible subsystems (IIS) of large mixed-integer programs (MIP) and bottlenecks in highly nonlinear networks. While detecting such errors is a significant achievement, removing them is extremely difficult.
In this presentation, we highlight our insights on data quality improvement. We report our results on data from the German high-pressure gas transport network using methods from data preprocessing and mathematical optimization.
EURO Practitioners’ Forum past and planned activities are available to the Forum members, as well as the wider public.
Visit the website and register as a member for free, to get the regular updates on all activities: EPF Member registration page. The recordings and details from previous webinars are also available on this website.
Follow the Forum on Twitter and LinkedIN , and feel free to get in touch.