Q-Learning- Sonic The Hedgehog
Abstract: NeuroEvolution of Augmenting
Topology (NEAT) technology has made advantages in the areas of solving complex problems in artificial intelligence such as video games like Sonic The Hedgehog.
Objective: To improve and design NEAT technology with the intentions of completing the first three levels of Sonic The Hedgehog (Green-Hill Zone) over time.
Methods:
Q-Learning is a model-free environment evaluated with the aim of maximised performance and efficiency. A python script uses q-learning for the Sonic The Hedgehog game imported from Steam. The value of current fitness variable is used to train the data.
Conclusions:
The higher the score, the more efficient the artificial intelligence will become in clearing the levels of the game.
