Abstract
A class of optimization algorithms for finding the global minimum of functions of continuous variables is presented. These algorithms merge conventional local minima search strategies with the stimulated annealing (SA) technique. The rationale behind these algorithms is discussed, and a complete description is given of one of them, derived from the Hooke and Jeeves (1961) search method. Tests made on mathematical functions show an increase up to two orders of magnitudes in efficiency with respect to a conventional SA algorithm. An example of application to VLSI design is given.

This publication has 8 references indexed in Scilit: