In fitness landscapes, starting from any genotype, change one allele randomly to go to neighboring genotype, and then stay there if it has higher fitness {adaptive walk}|. Adaptive walks go uphill to local peaks.
steps
After uphill steps, number of possibly higher steps becomes half or less and finding a higher step takes twice as long or longer.
If starting at low fitness, large steps work best. If starting at high fitness, small steps work best.
recombination
Recombination allows leaving local maximum to look for global maximum, if landscape is not too random.
few alleles
With two alleles for genes, number of steps to local peak is natural logarithm of gene number. Many local peaks are nearby. Because changes are random and neighboring fitnesses do not correlate, selection makes little improvement.
multiple gene changes
If several alleles change, steps go farther, and fitness-change range is greater.
gene number
If gene number increases, fitness increase per step is less.
Mathematical Sciences>Computer Science>Systems>Evolution Theories
3-Computer Science-Systems-Evolution Theories
Outline of Knowledge Database Home Page
Description of Outline of Knowledge Database
Date Modified: 2022.0224