- Dear all,
I have proposed a novel strategy for controlling the population diversity in genetic algorithms using binary representation.
I want to compare it with other up-to-date approaches designed to boost up the exploration capabilities of GAs like
- restart strategies,
- redundant representation, diploidy and multiploidy,
- self-adaptation of crossover and mutation probabilities
I am also seeking for a set of test problems on which the strengths and weaknesses of the approaches could be demonstrated.
Can anyone suggest any approach or benchmark to use?
> Can anyone suggest any approach or benchmark to use?I'd suggest have it go head to head in Chess or Checkers, Normal
> Jiri Kubalik
Genetic Programming on one side: You experament on the other side.
The one tactic that becomes the out right "champion" is the Best
tactic. You'd have to run it several times, but you will know the
answer. By the way, on chess if one of the "Test Subjects" uses more
the a certain amount of memmory it loses: to prevent it from crashing
your computer. Have it evolove a tree pruning algothrim.