Test Context

We use a canonical EA, as shown in figure 3. Each of the modules is self-contained and is explained below.

• Initialization Each gene in each chromosome is randomly initialized within its domain constraints.
• Evaluation Each chromosome is evaluated with a specified fitness function.
• New Population Generation The new population is formed in four steps:
  • Elitist. This is a common method of ensuring that the best chromosome from the last population is preserved intact. It is simply copied as the first member of the new population.
  • Crossover. We use tournament selection to produce the next tranche of population.
  • Mutation. We switch between parabolic search mutation and random mutation every iteration.
  • Re-initialization. Some of the new population is entirely re-randomized each iteration. This is normally a very low percentage of the chromosomes in the population.

 
Figure 3: Algorithm Context