Transformations enabling to construct limited-memory Broyden class methods
Jan Vlcek (vlcekcs.cas.cz)
Abstract: The Broyden class of quasi-Newton updates for inverse Hessian approximation are transformed to the formal BFGS update, which makes possible to generalize the well-known Nocedal method based on the Strang recurrences to the scaled limited-memory Broyden family, using the same number of stored vectors as for the limited-memory BFGS method. Two variants are given, the simpler of them does not require any additional matrix by vector multiplications. Numerical results indicate that this approach can save computational time.
Keywords: Unconstrained minimization, variable metric methods, limited-memory methods, Broyden class updates, numerical results.
Category 1: Nonlinear Optimization
Category 2: Nonlinear Optimization (Unconstrained Optimization )
Citation: Technical report No. V 1037, Institute of Computer Science, Pod Vodarenskou Vezi 2, 18207 Praha 8. December 2008
Entry Submitted: 02/25/2009
Modify/Update this entry
|Visitors||Authors||More about us||Links|
Search, Browse the Repository
Give us feedback
|Optimization Journals, Sites, Societies|