Linking Genetic Algorithms To Simulations

1
Linking Genetic Algorithms To Simulations
Activity Search
2
DEVS/GA Functionality
Generate digraph models with parameter values
from the initial population
Generate initial population of chromosomes
Use SimOptView to make one gene per parameter
in search space
Current Population of Chromosomes
Define mapping from genes bit sets to parameter
value ranges
Run simulation for each chromosome and obtain
fitness values
Transfer fitness values to chromosomes
Define mapping from simulation output to
fitness function taking account of optimization
objective
Apply genetic operators to get next generation
exit interpret best chromosome as parameter
values
3
GAs applied to Activity
4
GAs applied to Activity (contd)
5
GAs applied to Number of Maxima