📈 Strawberries are like OR, NVIDIA announcement and integrating algorithms in solvers, and will software continue leading the solver space?
Local Optimum: short, imperfect-yet-useful ideas - Edition #2
Welcome to a new edition of Local Optimum: a short, imperfect-yet-useful collection of ideas related to optimization, decision-making, and applied Operations Research.
Today’s edition will be about solvers…
Ready? Let’s dive in! 🪂
1) 🍓 Strawberries are like OR
Last week, I opened my fridge craving sweet strawberries… But they were awful: pale, sour, and deeply disappointing. Instead of throwing them away, I sprinkled some sugar, added milk, and left them in the fridge. Hours later, they had transformed completely: delicious, creamy, and sweet.
Operations Research often feels just like this. Your initial solution might seem disappointing, but through small, thoughtful adjustments (like a local search), you can turn those initial solutions into delight, local optimum ones. Next time your solution tastes sour, remember those strawberries: patience and careful improvement can make all the difference.
The strawberries now (delicious, btw):
2) 💡 NVIDIA announcement and integrating algorithms in solvers
If you weren’t in a cave recently, you might have seen NVIDIA's announcement about open-sourcing cuOpt.
It’s safe to say that the solver space will radically change in the coming years!
That same day, Ryan O’Neil shared his takeaways on solver developments after organizing the solvers cluster at the INFORMS Computing Society conference 👇
(click on the image below to go to the post)
A summary of the key takeaways:
Hybrid optimization is everywhere, and the power of some solvers like Google OR-Tools or Hexaly comes from combining multiple algorithms.
State-based modeling has a big opportunity, and thus Dynamic Programming will get a boost in the next years.
Established technologies are rapidly innovating, too, like using the Random-Key Optimization method jointly developed by Mauricio Resende (the author of GRASP) and other authors.
Exciting times ahead! 🚀
3) 🎯 Will software continue leading the solver space?
Not long ago I wrote about why software is the true driver behind faster solvers… And I believe that discussion is highly relevant to today’s post.
Key takeaways:
🚧 Hardware isn’t the golden ticket due to the inherent complexity of optimization problems, algorithmic bottlenecks, and the Amdahl’s law.
🧑🏻💻 Software is winning the race, and what took almost 30 days in 1988 took just 1 minute in 2003 using the very same hardware.
🔮 Software will continue to lead: emerging solvers and integration of Machine Learning promise further advancements.
If you want to read the full article, click here 👇
How do you crack the interview process?
In the last post I discussed how to create a good interview process from the interviewer’s perspective.
Next Monday, I’ll share the other part of the table: what to expect about interviews for OR Engineers, from the interviewee’s perspective.
I’ll cover:
📝 The application process
👀 Initial screenings and behavioral interviews
💻 Technical interviews, including possible questions and take-home assignments
If you’re looking for a job in OR, this will be useful. See you Monday!
And that’s it for today!
If you’re finding this newsletter valuable, consider doing any of these:
1) 📣 Partner with Feasible. I’m always looking for great products and services that I can recommend to subscribers. If you are interested in reaching an audience of Operations Research Engineers, you may want to advertise here. Just 📨 answer this email 📨 and I’ll get back to you.
2) 📤 Share the newsletter with a friend, and earn rewards in compensation.
If you have any comments or feedback, just respond to this email!
Have a nice day ahead ☀️
Borja.
The RKO method was truly interesting and promising.