Reproduction, genetic algorithms

Genetic Algorithm applied to the discussed data set leads to different subsets of selected variables. There are many different versions of GA, depending on the way reproduction, cross-over, etc., are performed. The algorithm used in our study, adapted from Leardi et al. [34,35], is particularly directed towards feature selection. In each GA run, a few subsets with similar responses are selected. Final solutions are then evaluated based on the RMSEP of an independent test set. Results are presented in Table 2 and in Figs. 5 and 6. 

The Genetic Algorithm (a) an initial population and the elimination of inefficient solutions (hollow circles), (b) generation of a new population through reproduction and mutation. 

Lately the genetic algorithm for minimization of the difference between simulated and experimental 1-D images was implemented this procedure allowed the best fit to be chosen automatically (64). A typical genetic algorithm (GA) consists of creation of the initial population, calculation of the fit to experimental data, selection of the couples, crossover (reproduction), and mutation. The approach and terminology are adopted from biology and resemble fundamental steps in evolution. 

Selection of the best members of the population is an important step in the genetic algorithm. Many different approaches can be used to rank individuals. In this example the ranking function is given. Highest rank has member number 6, and lowest rank has member number 3. Members with higher rank should have higher chances to reproduce. The probabihty of reproduction for each member can be obtained as a fraction of the sum of aU objective function values. This fraction is shown in the last column of Table 19.3. Note that to use this approach, our objective function should always be positive. If it is not, the proper normalization should be introduced at first. 

In order to illustrate a typical genetic algorithm in action, consider a population P of n entities represented as 8-bit codes as follows P = 11010110, 10010111, 01001001,. ... Then, suppose that at a certain point in the evolutionary process, the following pair of codes is selected to reproduce py 11000101 and pn = 01111001. Reproduction in this example is defined as a process whereby the couple exchanges the last three digits of their codes followed by a mutation process. In this case, the pair py and pn will exchange the last three digits of their codes as follows ... 

The principle may even be applied to the optimization of technical systems or material properties. So called genetic algorithms consist of a set of operators simulating reproduction, combination, and mutation applied to linear parameter sets defining the technical system (e.g. Goldberg (1989)). 

---