Genetic and Evolutionary Algorithms for PSP - Digital Signal Processing in Power System Protection and Control - page 275

Chr10 Chr9

Chr8

Chr1

Chr7

Chr6

Chr2

Chr5

Chr3

Chr4

trial 1 trial 3 trial 2 trial 6 trial 4 trial 6

1

2 3 4 5 6 7 8 9 10

0.82 0.95 1.0

0.0 0.18 0.34 0.49 0.62 0.73

NP =6 Selected: 1, 2, 4, 4, 6, 8

Fig. 13.3 Selection with the wheel of roulette method

pointer 1 pointer 2 pointer 3 pointer 4 pointer 5 pointer 6

1

2 3 4 5 6 7 8 9 10

0.82 0.95 1.0

0.0 0.18 0.34 0.49 0.62 0.73

random
number NP =6 Selected: 1, 2, 3, 4, 6, 9

Fig. 13.4 Selection with the stochastic universal sampling method

have higher chance to proceed to the next generation. One can see that both
''good'' and ''poor'' individuals can be selected; moreover, some of them can be
drawn many times. Here, after NP = 6 trials (wheel revolutions) the individuals 1,
2, 4, 4, 6 and 8 were selected.
The procedure of stochastic universal sampling is shown in Fig. 13.4 . Here, the
comb of pointers separated one from another by the same distance is used to
indicate the selected set of individuals. Its left-hand side position is determined by
drawing a random number. With this method the probability of repeated selection