# Cut search

### From Glossary

An algorithm strategy for
global optimization (notably for integer programming) that consists
of two alternating phases: *search* and *cut*. The search
phase finds linearly independent vectors emanating from a *root
point*, , to setup a probe for the cut phase. Usually, is
an extreme point, so the search is an *edge probe phase* that extends
edges of the cone rooted at until it intersects the candidate
solution points. Then, a cutting plane
is added (usually a convexity cut) to
exclude the root point. (A new root point is obtained by solving the new approximating problem on the smaller polyhedron.)