Experiments with a new selection criterion in a fast interval optimizationalgorithm

Usually, interval global optimization algorithms use local search methods t o obtain a good upper (lower) bound of the solution. These local methods ar e based on point evaluations. This paper investigates a new local search me thod based on interval analysis information and on a new selection criterio n to direct the search. When this new method is used alone, the guarantee t o obtain a global solution is lost. To maintain this guarantee, the new loc al search method can be incorporated to a standard interval GO algorithm, n ot only to find a good upper bound of the solution, but also to simultaneou sly carry out part of the work of the interval B&B algorithm. Moreover, the new method permits improvement of the guaranteed upper bound of the soluti on with the memory requirements established by the user. Thus, the user can avoid the possible memory problems arising in interval GO algorithms, main ly when derivative information is not used. The chance of reaching the glob al solution with this algorithm may depend on the established memory limita tions. The algorithm has been evaluated numerically using a wide set of tes t functions which includes easy and hard problems. The numerical results sh ow that it is possible to obtain accurate solutions for all the easy functi ons and also for the investigated hard problems.