Chapter 3: Heuristics
Society for Industrial and Applied Mathematics eBooks · 2013
The first section in this chapter is devoted to the presentation of some algorithmic schemes which can be used to find good solutions to GO problems. Methods included in this section presume that evaluating the objective function is sufficiently cheap—so cheap that it is often possible to perform local optimization runs. Thus most methods in this section will be based on the exploitation of local optimization and in the generation of samples of quite large cardinality. In the following section, models and methods are presented for GO problems in which evaluating the objective function is extremely expensive from a computational point of view. Thus the focus of Section 3.2 is on trying to save as much as possible in function evaluations, possibly devoting a large computational effort in deciding the few points where the objective function has to be evaluated. The last section in this chapter is devoted to some basic results concerning the capability of heuristic GO methods of eventually locating the global optimum.