Geometry Optimization and Conformational Analysis Through Generalized Simulated Annealing

Abstract

On statistical-mechanical grounds, a stochastic optimization technique (generalized simulated annealing) has been recently proposed which contains both classical simulated annealing (Kirkpatrick et al., 1983) and fast simulated annealing (Szu, 1986) as particular cases. This technique can be faster than both in detecting global (and also local) minima. Its utility in quantum chemistry is here illustrated, through the use of a semiempirical quantum method, on molecules of the series CH,-R (C,H,, CH,COH, CH,OH), H,X, (H,O,, H,S,), X,Y, (N,H4, P2H4, N,F,), for double bonds (C2H4 and CH,NH), and finally for H,O,. .