首页 > 代码库 > Evolutionary Computing: 5. Evolutionary Strategies(2)

Evolutionary Computing: 5. Evolutionary Strategies(2)

Resource: Introduction to Evolutionary Computing, A.E.Eliben


 Outline

  1. recombination
  2. parent selection
  3. survivor selection
  4. self-adaptation

 

1. Recombination

1.1 Basic recombination

two parents --> one child

There are two recombination variants distinguished by the manner of recombining parent allels:

  • discrete recombination: one of the parent alleles is randomly chosen with equal chance for either parents.
  • intermediate recombination: the values of parent alleles are averaged.

Formally, given two parent vectors x and y, one child z is created, where

技术分享

for all i ∈ {1,...,n}.

 

1.2 An extension

 An extension of this scheme allows the use of more than two recombinants, because the two parents x and y are drawn randomly for each position i ∈ {1,...,n} in the offspring anew.

These drawings take the whole population of μ individuals into consideration, and the result is a recombination operator with possibly more than two individuals contributing to the offspring.

 The exact number of parents, however, cannot be defined in advance. This multiparent variant is called global recombination. To make terminology unambiguous, the original variant is called local recombination.

 

 

 

2. Parent Selection

 

3. Survivor Selection

 

4. Self-adaptation

 

5. Example application

 

Evolutionary Computing: 5. Evolutionary Strategies(2)