Читать книгу Handbook on Intelligent Healthcare Analytics - Группа авторов - Страница 30

It emphasized its intrinsic malleability at various stages of definition with the introduction of GA. To explain this point further, we suggest an approach that differs considerably from the biological inspiration of GA and takes a method of child development that differs somewhat from gene exchange and crossover.

First, let us use the n-dimensional design area defined by a cartesian coordinate for two parents, created as described above and represented by points A and B. Now, draw a line between A and B and the location between A and B at that stage, according to the normal Gaussian middle point distribution, i.e., the midpoint is the most likely place. To begin with, construct a new n-dimensional coordinate system of Cartesian origin at O, whose coordinate axes parallel to the original system, extending from minus to further infinity. With a standard Gaussian distribution based on O along each axis, n coordinates can be produced that define the design point of children A and B. This stage might be dropped off the AB side. This will yield more than one child for a couple of parents and allows multiple likelihood distributions rather than gaussian.