11 Mar

This blog posts summarize the current questions on team-based mean-field theory. Mainly about: Gap between continuous portion of agents vs.. integer agents number Influence of numbers of agents Stationary distribution assumption Upper bound of k-hop neighbour error - exponential decay prop...

04 Feb

These slides are presented in group meeting. Talking about the evolution from individual Reinforcement Learning towards multi-agent mean-field algorithms. Especially, one of emphasis is that we discussed two different strategies based on mean-field. In order to understand easily, we named them age...

28 Jan

This improvement is based on the previous posts, corresponding to the first step: Step 1: Deal with information problem: using only closest neighborhood information. Code is here The main improvent is to set the observation field. Current mean-field assumes all agents' information are av...

20 Jan

Paper Target: IROS 2022 Current open source mean-field reinforcement learning code, which is based on MAgent Blue: Group 1 (Proposed Algorithm)Red: Group 2 (Opponent) Algorihtm Information Collision Task Current using all agents as neighborhood One state (position) only...

  • «
  • »