EURO 2025, celebrating the 50th anniversary of EURO, is one of the most significant events in the field of operational research and analytics. This conference brings together researchers, practitioners, and academics from around the world to discuss the latest developments, trends, and applications in operational research. The conference will feature a variety of presentations, workshops, and networking opportunities, providing a platform for participants to exchange ideas and collaborate on innovative solutions to complex problems.
Gurobi is proud to participate as a Premium Sponsor at EURO 2025. As leaders in mathematical optimization, we are keen to share our latest advancements and insights. Visit our booth and attend our expert-led sessions.
Speaking sessions at EURO2025
What’s new in Gurobi 12.0 – Michael Winkler (Monday 23 June 2.30pm – 4pm)
In this presentation, we will give an overview of recent developments within Gurobi 12. In particular, we will talk about performance improvements and the improved global MINLP solver.
Solving LPs using GPUs: a practical review – Dr. David Torres Sanchez (Tuesday 24 June 10.30am – 12pm)
In this talk, we will explore the reality of using GPUs to solve LPs from industry problems. Newly developed First Order methods (e.g., PDHG) and the Barrier algorithm can leverage this hardware for performance, but to what extent? And should we all rush to buy the new chips?
Generative AI for solving optimization problems: Is it helpful? – Dr. Mario Ruthmair (Tuesday 24 June 12.30pm – 2pm)
AI is reshaping many business fields, but can it reliably assist practitioners in solving optimization problems? This talk explores how Gurobi users can leverage Generative AI in their workflows. We discuss chat bots like the “Gurobot” and the “Gurobi AI Modeling Assistant” (both based on ChatGPT). Through examples, we will examine when AI enhances workflows and when human expertise remains essential.
From Infeasibility to Feasibility- Improvement Heuristics to Find First Feasibles for MIPs – Dr. Tobias Achterberg (Wednesday 25 June 2.30pm – 4pm)
Relaxation Induced Neighborhood Search (RINS) and other large neighborhood search (LNS) improvement heuristics for mixed integer programs (MIPs) explore some neighborhood around a given feasible solution to find other solutions with better objective value. This often leads to a chain of improving solutions with a high quality solution at its end, even if the starting solution is rather poor. RINS and its variants are the most important heuristic ingredients in Gurobi to find good solutions quickly. But they have one issue: they can only be employed after an initial feasible solution has been found. This initial feasible solution is usually found by other heuristics, socalled “start heuristics”, like rounding of LP solutions, fix-and-dive, or the Feasibility Pump. In this talk, we discuss a different approach, which works surprisingly well: similar in spirit to the Feasibility Pump, consider infeasible integral vectors as input to the improvement heuristics and search in the neighborhood for vectors with small violation to act as new starting point for the next LNS improvement heuristic invocation.
