Herd male genetic algorithm
H. John Caulfield · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2000
There are two extremes in optimization with little ground between them. They are local optimization and global optimization. Local optimization is normally very fast, but the optimum it finds may be far from the best (global) optimum. Global optimization is very slow, but it gives the best optimum- at least in principle. Using mammal herds as rough models, we suggest a new evolutionary method that has aspects of both and achieves intermediate results most of the time: faster-than-global convergence with better-than- local performance.