Heuristic strategy

Default value: normal
rosepy parameter name: heuristic_strategy

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

The heuristic_strategy parameter controls the application of primal heuristics during the MIP solving process. These heuristics attempt to find feasible solutions quickly without fully exploring the branch-and-bound tree. Early and effective heuristics can improve solver performance by providing good incumbent solutions to guide the search. Heuristics can be especially valuable for large or difficult problems where proving optimality is slow, but feasible solutions are still useful.

The available options are:

Minimal

rosepy parameter value: minimal

Applies only lightweight heuristics occasionally. This setting reduces the overhead from heuristic computation, making it well-suited for models where node throughput is prioritized over early feasibility.

Normal

rosepy parameter value: normal

A balanced strategy that applies a mix of commonly effective heuristics at appropriate stages of the search. Suitable for most problems, it offers a good trade-off between computation time and the likelihood of finding high-quality feasible solutions early.

Extensive

rosepy parameter value: extensive

Enables an aggressive use of heuristics, trying as many strategies as possible and at various stages. This approach is ideal when obtaining a feasible solution quickly is more important than solving time or when the model is difficult and benefits from strong primal guidance.