global search

Terms from Artificial Intelligence: humans at the heart of algorithms

Page numbers are for draft copy at present; they will be replaced with correct numbers when final book is formatted. Chapter numbers are correct and will not change now.

Global search is about finding the best overall region. of the fitness landscape. In the geographic analogy of finding the highest mountain, it is like getting to the Himalayas rather than the Serengeti, that is an overall area where one believes and optimal or at least very good soltion To do thos global search ahs to ignore or smooth out fine details of parameter choice. Global search is in contrast to local search, which is about finding the best point within a region. Simulated annealing is an example of an algorithm that starts off with a global search phase (when it has a high temerature) and then gradually moves towards local search.

Used in Chap. 9: pages 133, 134