A regularized limited memory BFGS method for nonconvex unconstrained minimization (Q2248965)

The authors propose a regularized limited memory BFGS method for solving nonconvex unconstrained optimization problems. The method is effective when the condition number of Hessian approximations tends to become ill-conditioned. The global convergence of the method is established when the Wolfe-type line search is adopted. The local R-linear convergence of the algorithm is analyzed. The effectiveness of the method is illustrated by some numerical results.