Limited Memory Gradient Methods for Unconstrained Optimization Problems
The limited memory steepest descent (LMSD) method stores a few past gradients to compute multiple stepsizes at once, providing an efficient approach for unconstrained optimization problems.
