r*****d 发帖数: 727 | 1 According to my understanding, Quasi-Newton is faster since this algorithm
does not check total energy and only follows the direction of the steepest
descent (force)? So it only works well when the intial structure is near the
minimum. In this way, is this Quansi-Newton method (IBRION=1) can be used
for Transition State optimization?
Another question, both of the two methods will find their instantaneous
groundstate. In other words, local minimum. So how to make sure we can find
global minumem? Increase POTIM? |