Project Summery

Within this project, our central aim is to harness the power of evolutionary algorithms in the training of neural networks, especially in situations where traditional data-driven approaches fall short due to limited data availability. These challenges frequently manifest in dynamic environments, such as gaming scenarios, where the constantly evolving nature of the data makes conventional training methods impractical.

To introduce the concept of neural evolution within the gaming context, we will embark on the development of a dedicated neural network. This network will be specifically designed to take critical decision-making parameters as input and produce corresponding outputs, effectively serving as a virtual controller, guiding the in-game agent towards achieving higher scores.