// the find
TimefoldAI/timefold-quickstarts
Get started with Timefold quickstarts here. Optimize the vehicle routing problem, employee rostering, task assignment, maintenance scheduling and other planning problems.
A collection of 15 runnable constraint-optimization examples built on Timefold Solver (a fork of OptaPlanner), covering scheduling and routing problems from VRP to flight crew assignment. Each quickstart is a self-contained Quarkus or Spring Boot app with a REST API and minimal UI. Aimed at Java developers who need to solve NP-hard planning problems without implementing metaheuristics from scratch.
The breadth of covered domains is genuinely useful — VRP, employee scheduling, conference scheduling, hospital bed allocation all in one place means you can find something close to your actual problem and adapt the constraint model rather than starting from scratch. The constraint provider pattern (ConstraintStreams API) is declarative and readable: you express what a good schedule looks like, not how to search for one. Each quickstart includes a working test for the constraint provider, which is the right place to catch logic errors before running a solver for 30 seconds. The solver itself (inherited from OptaPlanner) is mature and handles late acceptance, tabu search, and simulated annealing transparently.
These are demos, not templates — the domain models are hardcoded to the demo data shape, and pulling one into a real project means significant rework to connect your own persistence layer and data model. The Quarkus dependency is load-bearing and non-negotiable here; if you're on a different stack you're basically reading the constraint logic for inspiration only. Shadow variables and variable listeners, which several quickstarts use, are advanced Timefold concepts with almost no explanation in the code — you'll hit the solver docs hard before you understand what's happening in food-packaging or maintenance-scheduling. The fork notice at the bottom is fine legally, but Timefold also has a commercial cloud product they're steering you toward via the repeated 'off-the-shelf model' callouts in the README, so be clear-eyed that this is partly a sales funnel.