I write about my everyday life as optimization expert, where I translate business requirements to mathematical formulars, then to software -- and all the way back again.
How I left Vienna with cake (and a few lessons)
|
Hello Reader! Optimization is one of the most powerful technologies of our time — and yet, it rarely makes headlines. I'm currently returning from the Gurobi Summit 2025 in Vienna. Thus, this newsletter will be a little different: let’s start with three quotes worth highlighting from the event: My personal Top-3 of Quotes
“Two years from now, spam will be solved.” – Bill Gates, 2004 This quote was shown at the Summit, alongside many others from world-class figures. And it stuck with me. Not because spam is solved (we know it isn’t), but because it’s a perfect reminder that predictions are easy — and often wrong. What actually matters is not what we believe might happen tomorrow, but the actions we take today. In optimization, that means rolling up our sleeves and solving the problems right in front of us.
“We are super honest and accurate.” – Oliver Bastert, on why OR is less hyped than LLMs This one made me smile. Oliver has a point. And I truly hope this honesty always remains a defining feature of our field. In optimization, people rarely promise the moon without the math to back it up. Compare that with some corners of the AI world, where money circulates in ever larger circles, balance sheets inflate without real value created, and hype trumps substance. In contrast, optimization stands on applied math — arguably the most profound technological asset of the 21st century. Progress here is slower, less scalable than burning electricity and GPUs, but each step is real, lasting, and impactful. And yes, that’s me climbing back down from my soapbox ;)
“If you know what you’re doing and you can describe it, an optimization solver is the best tool you can ever use.” – Johannes Prescher, Head of IT & Tools Network Team, A1 Telekom I love this. It’s not from an academic researcher or a solver developer — it’s from someone in the trenches, working on real infrastructure problems. Johannes highlighted something important: in practice, the solver isn’t always used to prove optimality. His problem is way too big for that. Instead, the real value lies in generating good recommendations at the push of a button, under multiple objectives and constraints. And that brings me to one of my favorite talks from the Summit… Optimization in the WildJohannes’ presentation, “Automated Preliminary Planning and Cost Optimization in Fiber Rollouts – How Gurobi Accelerates Strategic Decisions”, pulled back the curtain on the invisible infrastructure that keeps us all connected. Most of us don’t think about fiber networks — but behind every modem and mobile antenna lies a complex web of copper and fiber connections. As Johannes explained, copper is fine for a few meters, but real capacity comes from fiber reaching all the way into neighborhoods and villages. Here’s where optimization shines:
The lesson? In industrial practice, optimality isn’t always the main goal. What matters is fast, flexible, scenario-driven decision support. That’s where optimization becomes a strategic asset. The SachertorteThe Summit also had a playful side: Gurobi organized a tuning challenge for all 250 participants. The task? Find non-default parameter values that improve solver runtime on well-known MIP benchmarks. The prize? A genuine Wiener Sachertorte — the iconic Viennese chocolate cake. I couldn’t resist. I tuned my set, submitted my run, and when the challenge closed, I had shaved the average runtime from 200 seconds down to just 45 seconds. To my delight, that meant I got to carry home the Sachertorte. For once, I won’t just return home from a “nerdy event” with ideas and stories — I’ll actually bring back something my family can taste. I’m really looking forward to sharing the Sachertorte with them. It might be the first time they get a sweet, tangible proof of what these optimization challenges can deliver. If you enjoyed these reflections from the Gurobi Summit, chances are someone in your network might appreciate them too. Feel free to forward this newsletter to a colleague who’s curious about optimization or simply enjoys a good story from the field, or tell them to subscribe :)
PS: The event venue was so grand, I learned Napoleon once proposed there. Optimization + history — not a bad mix. |
Bluebird Briefings
I write about my everyday life as optimization expert, where I translate business requirements to mathematical formulars, then to software -- and all the way back again.
Hello Reader! with wind and rain becoming ever-present, summer has officially given way to autumn. In this episode of Bluebird Briefings I'll give you a little summary of what happened in the world of optimization over the summer of 2025: Major Industry News Gurobi released the State of Mathematical Optimization 2025 report. It contains results from what's probably the biggest survey among optimization practitioners. The findings are impressive. They attest to the growth of optimization as...
Hello Reader! Every Superhero Has an Origin Story For GAMSPy’s machine learning superpowers, that story begins with a package called OMLT. OMLT is short for Optimization and Machine Learning Toolkit. It’s a package that extends Pyomo to integrate machine learning models directly into optimization problems. Because OMLT supports ONNX interoperability, you can take a model trained in TensorFlow, Keras, or PyTorch and embed it into a Pyomo optimization model. It also happens to be the package...
Hello Reader! Here is an intuitive description of the inner workings of 4 different approaches to solving optimization problems: 🔹 Heuristics: Fast & Practical A heuristic is a rule of thumb for decision-making. It doesn’t guarantee the best solution, but it’s often good enough. A strong heuristic has at least one of these qualities: ✅ It finds high-quality solutions quickly. ✅ It delivers decent results with minimal effort. Even if a heuristic fails often, it can still be valuable if it's...