The AID Method for Global Optimization
Mahamed G. H. Omran, Fred Glover · Algorithmic operations research · 2011
An Alternating Intensification/Diversification (AID) method is proposed to tackle global optimization problems, focu sing here on global function minimization over continuous varia bles. Our method is a local search procedure that is particul arly easy to implement, and can readily be embedded as a supportin g strategy within more sophisticated methods that make use of population-based designs. We perform computationaltests comparing the AID method to 20 other algorithms, many of them representing a similar or higher level of sophis tication, on a total of 28 benchmark functions. The results show that the new approach generally obtains good quality so lutions for unconstrained global optimization problems, suggesting the utility of its underlying notions and the pot ential value of exploiting its multiple avenues for general ization.