Menu
Home
People
Places
Arts
History
Plants & Animals
Science
Life & Culture
Technology
Reference.org
Bregman method
open-in-new
<p>The Bregman method is an <a href="/facts/Iterative_algorithm/uKMAJhEh">iterative algorithm</a> to solve certain <a href="/facts/Convex_optimization/7D6U4MTN">convex optimization</a> problems involving <a href="/facts/Regularization_(mathematics)/K601A3y2">regularization</a>. The original version is due to <a href="/facts/Lev_M._Bregman/T2JtKNhM">Lev M. Bregman</a>, who published it in 1967. </p><p>The algorithm is a row-action method accessing <a href="/facts/Constraint_function/QLFIFlFX">constraint functions</a> one by one and the method is particularly suited for large optimization problems where constraints can be efficiently enumerated. The algorithm works particularly well for <a href="/facts/Regularization_(mathematics)/K601A3y2">regularizers</a> such as the ℓ 1 {\displaystyle \ell _{1}} norm, where it converges very quickly because of an error-cancellation effect. </p>