Choose a supported problem family

Understand

The nine Modelers support Max-Cut, Knapsack, Custom QUBO, fixed-color Graph Coloring, fixed-cardinality Portfolio Selection, Set Cover, symmetric TSP, horizon-bounded Job Scheduling, and Capacitated Vehicle Routing.

TSP starts and ends at city 0. With three cities and symmetric distances, reversed tours use the same three undirected edges and necessarily tie. Four cities need nine encoded variables and exceed local execution.

Job Scheduling asks for feasibility within an integer horizon, not minimum makespan. Operations are non-preemptive, ordered within each job, and use fixed machines. Half-open intervals may touch without overlapping.

Vehicle Routing uses one depot, identical capacity-limited vehicles and positive customer demands. It serves each customer exactly once with at most the supplied fleet. Time windows, pickups, split deliveries and heterogeneous fleets are outside this implementation. Route diagrams are schematic; submitted distances are authoritative.

Portfolio Selection is a supplied mathematical objective, not investment advice. An encoded variable is not always a domain object: position, time and slack encodings can require several bits per object.

Visualize

Switch among all nine families to see how the same binary machinery is decoded into different domain views.

Experiment

Inspect a TSP reference tour, then a scheduling reference. Compare their domain objective labels. For routing, read the vehicle loads and the supplied capacity.

Return to the example controls