Skoči na vsebino

Ostalo & Traparije


KRUH

Recommended Posts

  • Odgovori 2,9k
  • Created
  • Zadnji odgovor

Top Posters In This Topic

Top Posters In This Topic

Popular Posts

Pozabil prilimat še mal RL vozikanja med prazniki. Po dveh letih sem se končno spet spravil in je bil že skrajni čas ker sem že fejst pogrešal in pozabil kolk fun je to ne glede na to kaj voziš ! Za

Posted Images

lol...

Danes bi mogu mojega mačka posnet, ko mu ni uspel enga ovinka spelat. Je prletu iz spalnice v vežo pol je pa hotu v dnevno sobo, pa mu ni znesl in je "zdrsnil" ven iz ovinka direkt v svoji posodi z briketi in vodo. Je vse strsel in polil. :lol:

...ma ne vem no kake ima 150cc 4t kart sile, če gre max 70kmh v ovink no, ne rečem da ga ni zanimivo vozit, sile pa glih nima velke...lahk pa da se za vse skup motim!

Tako je, motiš se! :wink:

Play, dej nehi sam govort, pa prid že 1x v Logatec, ko bomo vozil. Boš pol vidu kako je, ko greš na polno skozi desni ovinek pod mostom. :twisted: Mene vedno pol še par dni rebra na levi strani bolijo, da se komi premikam. Če pa primerjaš z Racelandom...tam bi pa tut jaz lahko med vožnjo še kkšen sendvič pojedu. [ok]

Link to post

Ful zanimivo. Avto se sam razvija na podlagi genetskega algoritma. Sam pustite da neki časa laufa. Na začetku je kr neki, pol se pa počasi začne razvijat v neki uporabnega

Še razlaga avtorja:

I was inspired to create this after seeing this implementation:

This version uses the Box2d physics library to make the car an arbitrary mass object and calculate the forces on it. It also uses color to show the evolution in progress.

It has 22 variables that represent 8 points of the car, the wheel size, and axle angles. The spring tension and torque are set according to the mass of the car.

It uses 2-point crossover and I'm not using bitstring representations and crossing over the entire real valued numbers at the cross points. Mutation simply replaces a particular value with another randomly chosen one. To keep it from converging too fast, it randomly chooses a mate sometimes and otherwise bases it on the fitness from the previous round.

I'd like to implement simulated binary crossover to more accurately represent the genetic search.

EDIT: black line = average fitness, red line = maximum fitness The number in parenthesis is the target score. Once it reaches that score it moves to the next car.

Povezava!

Link to post
  • 2 weeks later...

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Odgovori na to temo...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Povezava je bila samodejno vdelana.   Raje prikaži samo povezavo

×   Your previous content has been restored.   Počisti urejevalnik

×   You cannot paste images directly. Upload or insert images from URL.

  • Temo prebirajo   0 uporabnikov

    Te strani si ne ogleduje noben registriran uporabnik

×
×
  • Ustvari novo...