Recent research on neuroevolution has shown that it is a “competitive alternative to learning by gradient descent in reinforcement learning tasks.” (Hamon 2023). This project will focus on continued research into applications of neuroevolution in reinforcement learning tasks, using simulation environments to evaluate multi-agent systems. We shall note the different behaviors between agents, such as “predator-prey” or “leader-follower” relationships.
The simulations are displayed as 2D environments where either competitive or collaborative learning will be required, and agents can be implemented using Neuro-evolutionary models.
Gautier, Hamon & Nisioti, Eleni & Moulin-Frier, Clément. (2023). Eco-evolutionary Dynamics of Non-episodic Neuroevolution in Large Multi-agent Environments. 10.48550/arXiv.2302.09334.
Collaborate with Akbas to leverage artificial intelligence capabilities and develop an intuitive AI simulation platform that enhances user engagement and drives innovation across diverse applications.