Hello everyone! I ask if with Gdevelop it is possible to create a game that simulates the behavior of a robot, complete with energy progress bars and instantaneous speed indicators. The aim is to let him do research on a 2D map independently with the random generation of events. Do you think it’s possible?
In one word:
absolutely.
You could even make a pretty complex behavior tree to let him react to those events based on certain information or randomness.
Thanks so much. So I can also create a sort of intelligent behavior of the robot, in the sense that somehow he can learn (always in a very basic form)
well, not really.
But you can handle it like alian isolation does, by unlocking sertain behaviors, based on its experiences. So basicly, every time the bot has to make a decision, it runs throu its options and you lock certain options away until it sees/find “learned” something.
If you want to do something like AlphaStar behaviour powered by a deep neural network,
GDevelop is the wrong software (even tho i think it wouldbe theoretical possible).
Learning stuff in simple terms is doable, when its really simple.
Example recourse collection:
The robot get the task to collect a certain mineral. let him move around at random at first.
Everytime he stumbles opon some recourses, save the location of them and the amount.
Next time it has to go, it goes directly to the recourses, it already knows. a Learning progress.