Hi Uhrwerk,

the Selection-Algorithm i choosed in this example is the 'roulette wheel' sampling and is the simplest method. About the lifespan with my organism is true. they live only one generation. And yes this could happen in a loss of best agents. [ironic on] Thats life, every Olympic star will die some day :-) [ironic off]

Yes, the fitness-function compares the goal (here the modell-organism) to test for fitness-level. Usually goals for a serios problem space would be to find the most efficient, fewer task-timing spent or shortest waypath taken agent as a behavior.

Reminded, this should be the first simple example to beginn with AI. Extendable in every way to genetic programming, q-learning, reinforcment learning and all what evolutionary algorithm takes place.

Criticism is good in both ways negativ and/or positiv. This is how we got rewarded and could compare our actions and learn to do it better the next way. Words are enough but sometimes we need pain to learn, that the ofen is hot. hehe

i'm thankfull for your post.