References Geometry Optimization and Conformational 2-Dimensional case (D=2)

Concluding Remarks

We conclude from these preliminary studies that the MOPAC-GSA approach is a good qualitative and quantitative indicator of conformational molecular preference. We would like to emphasize that the GSA , differently from the gradient descent approach, enables us to map out local minima while the global minimum is searched.

We stress that this technique can be indifferently applied on all ''ab-initio'' or semi-empirical quantum methods, since the GSA routine makes no interference in the quantum calculus. In particular, we have used the semi-empirical MNDO-PM3 approximation, only for computational convenience.

The GSA method converges faster when the parameter q increases and has both the Classical Simulated Annealing (CSA) and the Fast Simulated Annealing (FSA) as particular cases. In this paper we have used (Cauchy machine or FSA) for both D=1 and D=2 cases. We have applied the algorithm in order to study a set of molecules which present one or more different equilibrium conformations by rotating a particular dihedral angle ( ) around the X-Y bonds. This procedure can be straightforwardly extended to any dimension D>2.

We acknowledge useful discussions with D.A. Stariolo. One of us (C.T.) also acknowledges the warm hospitality received from B. Widom at the Baker Laboratory, where part of this work was done.