Module genetic_algorithm::mutate

source ·
Expand description

The mutation strategy, very important for avoiding local optimum lock-in. But don’t overdo it, as it degenerates the population too much if overused. Use a mutation probability generally between 5% and 20%.

Structs§

  • Selects Chromosomes in the Population with the provided mutation_probability. Then mutates the selected chromosomes once using random mutation. The mutation probability is dynamically increased or decreased to achieve a target population uniformity
  • Selects Chromosomes in the Population with the provided mutation_probability. Repeatedly (number of rounds) mutates the selected chromosomes once using random mutation. The number of rounds is dynamically increased or decreased to achieve a target population uniformity
  • Selects Chromosomes in the Population with the provided mutation_probability. Then mutates the selected chromosomes once using random mutation.
  • Selects Chromosomes in the Population with the provided mutation_probability. Then mutates the selected chromosomes twice using random mutation. Useful when a single mutation would generally not lead to improvement, because the problem space behaves more like a UniqueGenotype where genes must be swapped (but the UniqueGenotype doesn’t map to the problem space well)

Enums§

Traits§