Review of Engelbrecht's Fundamentals of Computational Swarm Intelligence
- 594 Downloads
The search for excellent problem-solving algorithms is a kind of search like any other. It has its exploration and exploitation phases, its attractive local optima, its stochastic serendipitous surprises and foolhardy fixation on futile failures. As a distributed population of autonomous agents, the computational intelligence community follows the gradients or leaps from peak to peak to see what pays off. Some methods turn out not to be useful, some turn out to be varieties of something else, and some are found to be wholly new and powerful ways of doing things.
The swarm intelligence methods have largely grown up under the wing of the evolutionary computation paradigm. At first there was just a paper here and there at a conference; then there were enough to give them their own session, then several sessions. The link between EC and swarm intelligence may be a link of community without any crisp taxonomic rationality behind it; theory and methods pass among the various subdisciplines,...