is a simple example on how to optimize a multimodal function by using a canonical Genetic Algorithm implemented with JavaScript. The algorithm performs over the Function F6, which is formally defined as:
with the follow landscape:
By considering above information, optimize function F6 consists of finding values for x and y that archive the mininal score for z, i.e, f(x,y).
- Solution is represented as a binary vector with 44 bits, 22 for each variable x and y;
- One-point crossover;
- Bit flip mutation;
- Parents selection based on a tournament between k individuals;
- Merging population and offspring based on total ranking;
The page presents some parameters through which one may notice their impact in the algorithm behavior. Such parameters and their default values are:
- Population size: 100
- Mutation rate: 0.05
- Crossover rate: 0.80
- Maximal of generations: 200
- K-tournament: 3