Switch to full stage
Loading Project
Loading project …

Instructions

------------------------------------------------------------------------- AI Genetic Algorithm Pathfinding - #Machine learning ------------------------------------------------------------------------- Explanations : This is a AI using a Genetic Algorithm that is inspired of Drawin's Evolution theory(scroll to see more informations). Just watch the dots going further and further on the track. The top variable shows the generation number. The bottom variable shows the efficiency of the best solution. Turbo mode recommended ( ⇧ Shift and click the green flag) Instructions : Button Speed : Change the simulation speed Button "Optimized Mutations" : Ai will find the solution faster. Button Map : Change the Map (Map 9 : Create your track !) Button || : Play/Pause, See the best solution Run it here : https://turbowarp.org/522977651?turbo Genetic Algorithm : It is like the natural selection ! - The programm start by generating ramdom solutions. This will not work, the dot will crash. - The main loop : - SELECTION The programm selcect the best solutions and delete other bad solutions. - REPRODUCTION The solutions "create childrens" which are mix bettwen two parent solutions. - MUTATION There is a probability that the solution changes randomly --> The programm have created a new generation ! And the loop restarts. The solutions will become always better as you can see.

Notes and Credits

-------------------------------------------------------------------------Audio logo created by "Stereoalex" https://icons8.com/music/search/stereoalex ------------------------------------------------------------------------- Musique proposée par La Musique Libre Kubbi - Ember : https://youtu.be/7wQTcR3I44s​ Kubbi : https://soundcloud.com/kubbi​ -------------------------------------------------------------------------

TurboWarp is a Scratch mod that compiles projects to JavaScript to make them run really fast. Try it out by inputting a project ID or URL above or choosing a featured project below.