next up previous contents index
Next: Core algorithm Up: The Fib multimedia format Previous: Any complete Fib object   Contents   Index


The genetic algorithm

In this section, general design decisions and initial analysis for the genetic algorithm are established. The realized genetic algorithm is flexible and expandable designed.

The genetic algorithm is also an evolutionary algorithm. The term "genetic" refers to the ability of the algorithm to encode information of two or more individuals in a new individual and that it is working on information that code the (multimedia) object and do not represent those objects directly.

The algorithm is used to generate Fib objects, which represent a multimedia object as well as possible. The algorithm is given a particular multimedia object, for that it generates Fib objects /individuals, of which good will be selected. The creation of new individuals may also include analysis of the multimedia object and the use or analysis of information from other individuals. Which individuals are good, can be decided by the given parameters (by the evaluator for individuals).


The algorithm consists of five separate parts:

In Figure 8 a sketch for the flow diagram for the genetic algorithm is shown.

Figure 8: Flow diagram of the genetic algorithm
Image algorithmus

In the following the single Fib objects are called individuals. The set of all individuals, who are existing at a time in the algorithm, is calle population.



Subsections
next up previous contents index
Next: Core algorithm Up: The Fib multimedia format Previous: Any complete Fib object   Contents   Index
Betti Österholz 2013-02-13