Looking for "simple" dispatching software
Probably a long shot, but.... I am looking for software where I can plan routes, preferably free and/or open source, giant plus if it's a self hosted web based app, as I run Linux (I can potentially run in WINE, if needed) and other users run Windows.…
Probably a long shot, but.... I am looking for software where I can plan routes, preferably free and/or open source, giant plus if it's a self hosted web based app, as I run Linux (I can potentially run in WINE, if needed) and other users run Windows. Basically, I need to enter a list of addresses with time frames the driver needs to be at the location and time at the location, and it gives me the best route. If one driver can't do it, it provides the most optimized route for multiple drivers, up to the amount of drivers specified available. I've both used ChatGPT and personally searched for a solution and not found anything that would work. Even went as far as to set up Fleetbase, but it is quite a bit more than needed and is also pretty complex to work with. The list that's been suggested by ChatGPT is below. It provided lists of features, I only checked out the ones that looked like they may be usable, based on what it told me, and am also avoiding subscription based. Italics are what I've checked out. TIA! Fleetbase, VROOM, GraphHopper, OpenRouteService, Open Door Logistics, Studio Valhalla, LibreTaxi, Cabby, Detrack, RouteWork Traccar, OpenGTS, Geo2, Loop Route Planner, Ruvkgo, OptimoRoute, NextBillion(dot)ai Kaporeal TMS Onfleet, TrackRoad, Drivant, Shipday, RouteSavvy, RouteXL, PTV Map&Guide, RouteSmart
Collected discussion
The stack you're describing exists and it's free. It's just two pieces instead of one app. VROOM (vroom-project.org) is the solver. It handles exactly what you described: list of jobs, time windows, service duration at each stop, N vehicles, and it either fits them all or tells you what it couldn't place. Multi-vehicle, multi-depot, capacity constraints, all of it. C++, fast, Apache-licensed. VROOM doesn't know anything about roads, so you pair it with OSRM or Valhalla for the actual drive-time matrix. Both self-host fine on Linux. The fastest path: grab vroom-docker from the VROOM-Project org on GitHub. It bundles vroom-express (HTTP wrapper) and wires up OSRM for you. You download an OSM extract for your region from Geofabrik, let OSRM preprocess it once, and you have a REST endpoint you POST jobs and vehicles to. Runs on a small VPS or a spare box. Windows users just hit the web endpoint, so your mixed OS situation is a non-issue. The catch: VROOM is an API, not a product. There's a basic demo frontend in the repo that will get you a map and a route, but it's not a dispatch UI. If you want something with a real UI out of the box, look at Timefold (the OptaPlanner successor). Their vehicle routing quickstart ships with a REST API and a working web UI, Apache 2.0, and it does VRP with time windows. Java or Python. More code to touch than VROOM, but you get a running app on day one. Skip OR-Tools unless you enjoy writing solver code. It's excellent and it's free, but you're building the whole thing yourself. Fleetbase being too heavy makes sense. It's a whole fleet management platform and you want a solver. Now the part nobody mentions until you're two weeks in: The optimizer is the easy part. What kills real routing is upstream of it. Bad geocodes on new construction and apartment complexes will put a stop three miles from where the driver actually needs to be, and the solver will confidently build a beautiful route around a wrong pin. Second, service time per stop. If you feed it a flat 10 minutes and your real stops run 5 to 45, your afternoon time windows are fiction by 11am. Track actual dwell time per stop type before you trust any output. Third, the plan is only correct at the moment you generate it. Free tooling optimizes once, in the morning. It won't re-sequence when stop 4 takes an hour or the customer isn't home, and that's when most of the day's damage happens. None of that is a reason not to do this. Just build the habit of measuring planned vs actual on every stop from day one, or you'll never know whether the solver is wrong or your inputs are. Disclosure so nobody has to ask: I run a logistics software company, so I'm not a neutral party. I'm not pitching you. You asked for free and self-hosted and that's a legitimate ask, so that's what I answered.
Wow. That is quite a bit of great information to get me going. Thank you so much for taking the time. I know what I'll be doing the next few days 🙂
That's a very tall order, basically asking for 3 layers - the route optimization aspect, the distance/time aspect, and then the timeframe aspect. I can't think of a free solution that does all three of these things, but MapQuest will do the first and second - you'll get an optimized route with a good estimate of how long it will take, and then from there you can reason out if you will need multiple drivers. Only up to 26 stops though. Beyond that, you're looking at paid options.