There are two different practical ways of global path-following (PF) on pot
ential energy surfaces (PESs) of molecular reactions: (i) the Elber-Karplus
(EK) method (and its improvements), and (ii) the family of DDRP methods. T
he early versions of the methods under (i) are based on minimizing a functi
onal of the entire path and applying penalty functions as constraints, and
are evaluated only for molecular mechanical (MM) PESs. The first true impro
vement in the EK sequels is the method of Chiu et al. who - instead of usin
g penalty functions - introduce a redistribution of the grid points to subs
titute the constraints employed in former versions of the EK method, and pe
rform PF on quantum mechanical (QM) PESs. In the present paper the mathemat
ical foundations and the performances of the above methods have been compar
ed. The superiority of the DDRP method in accuracy and stability over the o
ther methods has been verified and tested by a difficult mathematical funct
ion simulating the conformational change in the catechol molecule. (C) 2000
Elsevier Science B.V. All rights reserved.