RoutingLab

Groningen has 1,120 underground waste containers that need to be emptied before they fill up. Every deposit is registered with an access card, but volume and weight are not measured. So the municipality must decide which containers to empty from deposit counts alone. Today that happens in two steps. They first select 250 clusters that are expected to fill up soonest, and then route through the selected clusters. The selection step is blind to routing cost, so trucks detour to far-out clusters while driving right past convenient ones that were never selected. We built a policy that makes both decisions at once. Deposit counts are used to estimate overflow probabilities that are used as prizes, and a single prize-collecting routing problem is solved using PyVRP to decide which clusters to empty today and in what order. Our simulations based on a full quarter of real deposit data (~2 million deposits) show: - 41% less driving, 251 km down to 146 km per day - Route durations of 3.8 hours instead of 6.8 - Same 99.9% service level - Containers emptied at 80% full instead of 49% Read more about the case study here: https://lnkd.in/ebWxHScv. These routing problems can be solved with PyVRP in under a minute, using a single CPU core. Interested in learning more? Get in touch via https://routinglab.tech/!

We're now offering consulting and implementation services at RoutingLab. This includes integrating and extending PyVRP for your exact use case or building custom optimisation algorithms for non-standard routing problems. A few areas we've worked in: ♻️ Sensor-based routing for municipal waste collection 🛒 Simulation models for same-day grocery delivery 📦 Integrated order picking and vehicle routing Get in touch at https://routinglab.tech if you need help with your challenging route optimisation problems.

𝐒𝐨𝐥𝐯𝐞 𝐲𝐨𝐮𝐫 𝐫𝐨𝐮𝐭𝐢𝐧𝐠 𝐩𝐫𝐨𝐛𝐥𝐞𝐦𝐬 𝐢𝐧 𝐦𝐢𝐧𝐮𝐭𝐞𝐬 𝐰𝐢𝐭𝐡 𝐏𝐲𝐕𝐑𝐏 𝐄𝐧𝐭𝐞𝐫𝐩𝐫𝐢𝐬𝐞 We're excited to announce PyVRP Enterprise, packed with additional routing features extending open-source PyVRP. Available today: ☕ Break planning respecting EU/US hours-of-service regulations 🔄 Reloading at depots 🚛 Vehicle-stop restrictions On the roadmap: 📦 Shipments with paired pickup and delivery stops 💡 Explainability ("why is a stop unplanned?") ⚖️ Fairness through balancing load, duration and distance across drivers 🗺️ Hard and soft clustering preferences ⏰ Time-dependent travel durations for traffic-aware routing Available through annual licensing as a pre-compiled Python package or with full source-code access. Early adopters benefit from extensive support and roadmap influence. Reach out to us at https://routinglab.tech.

𝐑𝐨𝐮𝐭𝐢𝐧𝐠𝐋𝐚𝐛 𝐏𝐥𝐚𝐧𝐧𝐢𝐧𝐠 𝐀𝐠𝐞𝐧𝐭 We're excited to announce the launch of our new route planning #agent, powered by the FastVRP route optimisation API. Our route planning agent works like a co-pilot you can talk to. Instead of manually building routes by clicking through complex interfaces, you simply describe what you need in plain text: → "Optimise for lowest cost" → "What if we add 10 more stops?" → "Should we hire another driver?" The agent instantly evaluates the #routing scenarios, explains trade-offs, and provides follow-up suggestions. This takes planning from hours to minutes, with just a single conversation. A day of planning work. Done in minutes. Book your demo now at https://routinglab.tech.

We have spent the past few weeks diving deep into EU and US regulations on #driver #break planning. Based on that reading we have now implemented fully compliant EU and US break planning for work and drive time regulations inside our routing solver. Ready to optimise your #routing? Get in touch at https://routinglab.tech/.

𝗪𝗲 𝗮𝗿𝗲 𝗵𝗮𝗽𝗽𝘆 𝘁𝗼 𝗮𝗻𝗻𝗼𝘂𝗻𝗰𝗲 𝗻𝗲𝘄 𝗙𝗮𝘀𝘁𝗩𝗥𝗣 𝗳𝗲𝗮𝘁𝘂𝗿𝗲𝘀: 🧩 Vehicle-task restrictions. You can now specify per task which vehicle(s) must be used to service that task, or which vehicles must absolutely not be used. This lets you assign drivers to familiar areas and tasks, or can be used to ensure the right vehicles are used for the right task. ⏰ Setup durations at new locations. These can for example be used to model the time it takes to park once at a building before delivering a number of parcels there. Setup is incurred only once per location. 🗺️ Custom routing profiles, which can be used to specify your own distances and durations for vehicle types FastVRP does not yet know about. 🪝 Webhooks for job status changes and real-time solve progress updates. This lets you observe newly found solutions in real-time during the optimisation run. Ready to optimise your #routing with FastVRP? Get in touch at https://routinglab.tech/.

𝗙𝗮𝘀𝘁𝗩𝗥𝗣 𝗔𝗣𝗜 𝗶𝘀 𝗻𝗼𝘄 𝗴𝗲𝗻𝗲𝗿𝗮𝗹𝗹𝘆 𝗮𝘃𝗮𝗶𝗹𝗮𝗯𝗹𝗲! FastVRP solves complex vehicle routing problems within minutes. Included features: ⚡ 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲: Route thousands of tasks in minutes. 🗺️ 𝗘𝗮𝘀𝘆 𝗶𝗻𝘁𝗲𝗴𝗿𝗮𝘁𝗶𝗼𝗻: Built-in integration with OpenStreetMap. ⏰ 𝗧𝗶𝗺𝗲 𝘄𝗶𝗻𝗱𝗼𝘄𝘀: Routes that respect your time windows. 🧩 𝗦𝗸𝗶𝗹𝗹𝘀: Handle task and vehicle skill compatibilities. ⚖️ 𝗣𝗿𝗶𝗼𝗿𝗶𝘁𝗶𝗲𝘀: Balance urgent and regular tasks. 🔄 𝗥𝗲𝗹𝗼𝗮𝗱 Reload at depots and continue delivering. Ready to optimise your #routing with FastVRP? Get in touch at https://routinglab.tech/.

𝗣𝘆𝗩𝗥𝗣 𝘃𝟬.𝟭𝟯 𝗿𝗲𝗹𝗲𝗮𝘀𝗲: 𝗢𝗽𝗲𝗻-𝘀𝗼𝘂𝗿𝗰𝗲, 𝘀𝘁𝗮𝘁𝗲-𝗼𝗳-𝘁𝗵𝗲-𝗮𝗿𝘁 𝗩𝗥𝗣 𝘀𝗼𝗹𝘃𝗲𝗿 Happy new year! We are excited to announce the release of PyVRP version 0.13, which adds support for: 🚀 Warm-starting from initial solutions; 🚚 Depot loading times; ⚡ Better performance on larger instances. We did a complete overhaul of the internal algorithmics by replacing our hybrid genetic search implementation with an iterated local search. This enables PyVRP to solve larger instances much faster than before, as can be seen in the figure below, benchmarking against the previous release on the well-known Gehring and Homberger instances with a runtime of just one minute and one core. PyVRP version 0.13 is now available on the Python package index and we have also updated FastVRP, our routing API, to use it. Curious about FastVRP? Please consider joining our wait list at https://routinglab.tech/ to be notified when we leave beta. Do not hesitate to contact us if you want to learn more about RoutingLab's products and services!

𝗙𝗮𝘀𝘁𝗩𝗥𝗣 - 𝗡𝗲𝘅𝘁-𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 𝗥𝗼𝘂𝘁𝗲 𝗢𝗽𝘁𝗶𝗺𝗶𝘀𝗮𝘁𝗶𝗼𝗻 𝗔𝗣𝗜 We're excited to announce the launch of FastVRP - a next-generation route optimisation API built by RoutingLab. Over the past three years, we've been building PyVRP, our open-source vehicle routing problem solver. Now, we're taking the next step by bringing its state-of-the-art capabilities to a fully featured API. Built for real-world vehicle routing applications, FastVRP offers high-performance route optimisation through an easy-to-use and production-ready API. FastVRP directly integrates with distance and travel time data from OpenStreetMap, and we also provide dedicated support to solve your complex routing problems - so you can deliver to your customers faster! Interested? Join the waitlist now at https://routinglab.tech or have a look at https://lnkd.in/e4j4htj7 to see what FastVRP can do. We’ll be gradually providing access to users as we continue to extend FastVRP to solve more complex routing problems. Stay tuned!