Chapter 3 . Applications of Genetic Aigorithms
Summary of Genetic Evolution
Randomly Create Population of Solutions
\--Select Parent Individuals
...
.}
•
•
Produce Children (Crossover and Mutation)
1
tf· ~~
Replace unfit individuals ~"('
Continue until one individual meets success criterial
(e.g., all EC50s predicted within 2X). Evaluate on your
testing dataset to assure you didn't over train.
39
Figure 3.2. Summary of the process of evolution with a genetie algorithm. The
population of size n is randomly created where each individual is a chromosome
constructed of genes (e.g., genetic material). Parents are selected from the
population and reproduce to create children. The children undergo mutation and
crossover. Their fitness is evaluated and the unfit individuals in the population are
replaeed. The process continues until a desirable outcome is aehieved.
33
29
~ 25
..
oe 21
~
17
13
0
50000
100000
150000
200000
Generation
Figure 3.3. Example of the variability in fitness scenarios between different
genetic algorithm simulations using the same parameter set. The genetie
algorithms was run 10 times using the same set of parameters. The difference in
fitness scenarios is a result of the random initialization of the population and the
random nature of the mutation and eros so ver operators.