Cut strategy

Default value: normal
rosepy parameter name: cut_strategy

See the linked article for information on How to change and view the solver parameters.

The cut strategy parameter controls the application of cutting planes during the solution of mixed-integer programming (MIP) models. Cutting planes (or "cuts") are additional constraints that tighten the linear relaxation of the problem, often helping the solver converge faster by reducing the size of the branch-and-bound tree. Choosing an appropriate cut strategy can significantly impact solver performance, depending on problem structure and solver settings.

The available options are:

Minimal

rosepy parameter value: minimal

Enables a lightweight cut strategy that prioritizes node throughput over cut generation. Only a small number of inexpensive and high-impact cuts are applied, reducing overhead and maintaining a fast pace through the branch-and-bound tree.

Normal

rosepy parameter value: normal

Applies a balanced mix of cut types. This is the default strategy, aiming to improve LP relaxations without significantly slowing down per-node solve time.

Aggressive

rosepy parameter value: aggressive

Enables full cut generation, using as many cuts as possible. This strategy often produces a tighter relaxation and can reduce the number of branch-and-bound nodes, but may increase the time spent per node. Best suited for hard problems where early pruning of the search tree is more beneficial than fast per-node evaluation.

Off

rosepy parameter value: off

Disables all cut generation. The solver will rely purely on branching and heuristics. This mode can be useful for benchmarking or diagnosing the impact of cuts on performance or solution quality.