Unlocking OR for all: from Models to Business Impact

Sylwester Ciodyk, Ortec Poland

Abstract: In today’s fast-moving world, optimizing efficiency and effectiveness is essential for every company. Drawing on my experience designing and deploying optimization systems at ORTEC, I’ll share field-tested stories of how we connected algorithms to human goals and operational constraints. Join me as we reveal how optimization drives business transformation, with real-life examples such as smarter load building in e-commerce and improved forecasting in the energy sector, leading to significant operational savings. Expect pragmatic takeaways you can apply immediately—how to quantify impact, handle uncertainty, and bridge the gap from prototype to production—to deliver measurable savings and better service.

Biography: With 15 years in Product and Project Management, Sylwester specializes in optimizing processes and driving efficiency across Automotive, Location Intelligence, Energy, and Retail. Currently, supporting the Director of Operations for the European Energy sector, Sylwester oversees key performance indicators, assists executive leadership, and leads initiatives focused on cost-effective solutions and automation. Known for combining strategic insight with practical execution, Sylwester helps organizations achieve measurable efficiency and growth.

Paweł Lichocki, Google

Abstract: In this talk, we present Mixed Integer Programming (MIP) from an informal, practitioner’s perspective. MIP is a powerful, declarative approach to solving combinatorial optimization problems. One expresses the problem at hand via a mathematical model, while a “black-box” solver does the number crunching and finds a solution. In practice, building a MIP-based application entails a few challenges. The optimization model must faithfully capture the reality of the problem while remaining “solvable” by the available solvers. Moreover, the optimization component must be integrated with the production system in a clear and robust way. We describe these issues in examples distilled from real-life applications of MIP to large-scale computer systems. We discuss the art of crafting efficient formulations for typical requirements observed in computer systems, like fault tolerance or churn control. We provide practical hints on how to encapsulate and verify mathematical optimization within a larger production environment. Finally, we discuss the relevant engineering challenges such as input/output validation, error resilience, and the ease of adding future extensions. Overall, we hope to shed light on the sometimes overlooked practical aspects of applying principled mathematical optimization to live and large-scale systems.

Biography: Pawel Lichocki is a software engineer at Google in the Operations Research team. For the past +10 years, he has been contributing to dozens of projects relying on combinatorial optimization. In particular, he applied Mixed Integer Programming in many real-life, large-scale systems and products. His interests include both practical and theoretical aspects of optimization, as well as evolutionary biology, meta-heuristics, and machine learning. He received a PhD from the School of Computer and Communication Sciences at EPFL in 2013 for work on the evolution of division of labor in multi-agent systems. Prior, he was a researcher in the Supercomputer and Networking Center in Pozna,ń where he worked on parallel and distributed processing algorithms. Recently, he has been exploring the possibility of using machine learning to improve MIP solvers.

Michał Kłos, PSE Polish Transmission Operator

Abstract: While Operational Research (OR) traditionally focuses on cost optimization, real-world applications often demand more than a single dimension of the objective function. In many real complex systems, the concept of “fairness” emerges as a crucial, yet subjective, element of the objective function. This talk explores how OR can be used to model and achieve equitable outcomes, even when “fairness” lacks a universal definition and must be translated into quantifiable objectives. We argue that incorporating fairness requires a deliberate shift from purely technical optimization toward a more nuanced approach that integrates stakeholders’ perceptions to equity considerations.

We present two distinct case studies from the energy sector that illustrate this challenge: Cost Distribution of Congestion Income and Cost Sharing of Remedial Actions. Through these examples, we demonstrate that achieving a “fair” outcome is not an add-on but an integral part of the modeling process. We conclude that successful applications of mathematical programming in this domain require not only technical expertise but also a deep understanding of the human and policy factors that define what is perceived as a just and equitable solution.

Biography: Michał Kłos is the Deputy Director for Computational Method Development at the NCAE Bureau (Polskie Sieci Elektroenergetyczne S.A., the Polish TSO) and former Director of the Interdisciplinary Center for Energy Analyses (ICAE) in the National Centre for Nuclear Research. He specializes in mathematical modeling and optimization for energy markets and power systems, with a focus on implementing EU regulations.

His work involves developing practical algorithms, including the clustering algorithm used in the first ENTSO-E Bidding Zone Review. He also co-authored the IP-DSR service concept, being responsible for the contracting and operational dispatch algorithms for service providers. His expertise includes leading energy market modeling projects for Transmission System Operators and acting as an expert on cost-sharing methodologies and congestion income distribution.

A graduate of Computer Physics from Jagiellonian University, he has authored over a dozen scientific publications and is a regular speaker at the annual European Energy Markets conference.

Geoffrey De Smet, Timefold

Abstract: Almost 20 years ago, I created an open source solver, now called Timefold Solver. These days, it is used across the globe to optimize vehicle routing problems, job shop scheduling, shift rostering and many other planning cases, directly or indirectly affecting the lives of millions of people.
In this keynote, I’ll cover the challenges, successes and lessons learnt. What makes our solver unique. How our engineering culture handles trade-off decisions. And why the timing is right for all of us to make Operations Research mainstream.

Biography: Geoffrey De Smet is founder/CTO of Timefold and the creator of OptaPlanner, its predecessor. He’s an expert in planning optimization, an active open source contributor and an international speaker. His technology is in use across the globe, to automate complex scheduling and routing of hundreds of thousands of employees, vehicles and machinery every day. Before founding Timefold, he worked as a Senior Principal Software Engineer at Red Hat (IBM), project lead at Schaubroeck and as an academic researcher at Kaho Sint-Lieven (KU Leuven).

Tomasz Stopa, IBM

Abstract: For over 40 years the universal quantum computer was a holy grail for scientists. Today, we already have noisy, intermediate scale quantum computers (NISQ) to play with. These current devices are already believed to allow getting results beyond reach for classical supercomputers.

In the talk we’ll explore the current state of quantum computing technology on the example of IBM, we’ll see how to play with these toys and why they should become part of your scientific explorations. The talk will provide a review of promising OR applications of quantum computers and algorithms to start experimenting with. We’ll also explore current limitations and necessary gain of capabilities to allow for tackling really meaningful OR problems. Finally, we’ll cover when we plan to get the actual holy grail (error corrected quantum computer) and how we believe it will look like.

Biography: Tomasz Stopa works as software development manager in IBM’s Kraków Lab, Poland. He obtained his Ph.D. in theoretical solid state physics from AGH University of Science and Technology.

IBM Master Inventor with 35 patents and scientific publications. As IBM Quantum Ambassador promotes quantum computing within industry, academia and high schools. Co-organizer of Kraków Quantum Computing Seminar (KQIS) and mentor in quantum computing hackathons.