Abstract
A stochastic self-regulating simulated annealing optimization method is presented, and compared to other optimization methods such as the simplex, steepest descent, and the recently proposed fast fitting method by Penna [Phys. Rev. E 51, R1 (1995)]. The presented method converges faster towards an acceptable set of optimization parameters than the other methods, and it is less susceptible to local minima of nonconvex functions. Examples are shown for fitting a simple two parameter Gaussian function and a complicated multiple parameter three-body interaction potential function.

This publication has 5 references indexed in Scilit: