M2M Machine-to-Machine

Combinatorial optimization problems arise across logistics, scheduling, and engineering, and C++ remains a language of choice when performance matters. This lecture takes a practical look at solving such problems in C++, using the Electric Vehicle Routing Problem as a running example. We begin with an exact solver, formulating the problem as a mixed-integer program and trying to solve it through GLPK's C API directly from C++. Exact methods, however, quickly hit their limits on real-world instances. The second half of the talk turns to metaheuristics: how they are designed, why C++ is particularly well suited for implementing them, and what design choices matter most in practice. We'll walk through a concrete implementation, touching on data structures for fast neighborhood evaluation, generic algorithm design with templates, and the performance considerations that separate a prototype from a production-ready solver. The goal is not to advocate for one approach over another, but to show how C++ supports the full spectrum of optimization techniques. # About Speaker Luka Matijević is a researcher at the Mathematical Institute of the Serbian Academy of Sciences and Arts, whose work bridges academic research and freelance engineering. He's drawn to identifying the right algorithm for a given problem — and to the kind of careful optimization captured by one of his favourite Futurama lines: *"When you do things right, people won't be sure you've done anything at all."* ## 📅 Event Details | | | | --- | --- | | 👤 **Speaker** | **[Luka Matijević](https://www.linkedin.com/in/luka-matijevic/)** | | 🕕 **Date & Time** | **27th of May (Wednesday), 6 pm** | | 📍 **Location** | **Beograđanka, Vidikovac, 22nd floor** | | 🏢 **Address** | **Masarikova 5** | | 💻 **Online** | **[C++ Serbia YouTube](https://www.youtube.com/@cppserbia)** | See you!

**Important time note:** Please plan on arriving between 5:30 and 6:00 as the elevators lock after 6 and you'll need to message us and we'll need to come get you. The building address is 4450 Bridge Park The entrance is 6620 Mooney St, Suite 400 You will need to scan your ID at the door to get a visitor badge. **Abstract** Like it or not, AI agents are now capable of turning a quickly written paragraph of requirements into a pull request that is ready to be integrated into real-world production applications and it's now our responsibility to make sure AI doesn't go rogue and take down prod - or corrupt our data by misunderstanding the requirements or our existing schemas. In this session we'll explore strategies to protect our codebases through unit and integration testing, documentation, and code review along with additional ways of providing context and guard rails to our AI agents as they carry out the work we've assigned them to do. By the time we're done, you'll have a firm grasp of the problem and understand some helpful options for protecting your codebase from vibe coding mishaps getting YOLOed into prod. **YouTube Link** TBD

Most people use AI like a search box: they type one sentence, hope for the best, and get frustrated when the answer is generic, wrong, or useless. But getting better results from tools like ChatGPT, Claude, Gemini, and other AI systems is not about memorizing magic prompts. It is about learning how to give the AI better context. In this beginner-friendly session, we’ll break down how to make AI dramatically more useful by improving the way you communicate with it. You’ll learn how to give clearer instructions, provide examples, set constraints, ask for better output formats, and use follow-up questions to turn a mediocre answer into a genuinely useful one. We’ll cover practical techniques you can use immediately for work, learning, writing, coding, planning, research, and everyday problem solving. We’ll also touch on why these same ideas show up in more advanced AI systems, including RAG, agents, evaluations, and AI workflows. No technical background required. Bring your curiosity, your questions, and maybe one real task you wish AI was better at helping you with. **What you’ll learn:** * Why “better prompting” is really about better context * How to structure requests so AI gives more useful answers * How to use examples, constraints, and output formats * How to iterate when the first answer is not good enough * How these skills connect to more advanced AI workflows This meetup is for anyone who wants to move beyond basic ChatGPT usage and start getting more practical value out of AI. LOGISTICS AND PARKING: The talk starts at 7:00 PM. The first half hour is reserved for everyone to get set up and mingle. Free pizza and drinks! The cheapest parking option is to find street parking, which will only cost you a few bucks. Otherwise, park in the nearby veteran's museum lot for $8. It's highly recommended you avoid the nearby $15 garage parking.

Couldn't make Knowledge? Or want to debrief the dancing robots or Autonomous Workforce? Join us for a Knowledge '26 recap, insights sharing, a couple rounds of golf, and fun at Topgolf. As always, drinks and food will be provided. More details to come — mark your calendar, RSVP, and save your spot today!