No description, website, or topics provided.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
Executable JAR


This project shows that MLP (Multi layer perceptrons) can be used in videogames to create characters with advanced AI capabilities. The main advantage of this aproach is the ability that neural networks have to simulate the natural learning process.

In this example, a neural network is used to play the clasical Lunar Lander game, and trained by means of a Genetic Algorithm that optimizes the MLP performance acording to the fitnes function. The fitness function penalizes fuel consumption and impact velocity against the floor.

This architecture makes possible creating games where AI characters evolve, making the game harder as the human player acquires new skills.

NetBeans project is provided in addition to the executable JAR. The following libraries are used:

Further explanations can be found here (in spanish):