next up previous contents index
Next: Analogy to the natural Up: The genetic algorithm Previous: Why genetic algorithms are   Contents   Index


Complexity estimation

With the approach of section 12.1 on page [*], it is possible to converted an arbitrary raster image into an Fib object, with a computing time growing linearly to the number of points and ther properties, as each point and its properties can be easiely added to a list element and the required time for the creation of the other Fib elements (of the root-element) can be performed in constant time.

This approach can be extended to any multimedia data, which are represented as properties of points of a finite, euclidean and discrete space. Thus always an operator can be constructed, that converts a multimedia object in linear time with the number of points and properties into a corresponding Fib object. This Fib object grows only linearly in size with the number of points and ther properties of the original multimedia object. This operator, however, doesn't produced good Fib object representations, because it does not use the possibilities of Fib.

How much effort is required to produce better Fib objects, is very difficult to estimate, since both the operators and the multimedia objects can be arbitrary. However, it can be assumed that, for multimedia objects with simpler structures the effort is lower.


next up previous contents index
Next: Analogy to the natural Up: The genetic algorithm Previous: Why genetic algorithms are   Contents   Index
Betti Österholz 2013-02-13