balajikatyan/b-ace

B-ACE is a Godot-based beyond-visual-range air-combat simulator with a Python PettingZoo interface for single- and multi-agent reinforcement learning. The main workflow is a Python agent or training script configuring and resetting the wrapper, exchanging actions and observations with the Godot simulation, and using the resulting rewards and episode outcomes to train or evaluate policies. The README documents this interaction; sampled Python files confirm the wrapper’s Godot transport and training integration. Some Godot internals and optional example integrations are represented at subsystem level because their implementations were not sampled.