Games

Set-up: Two-player game where agents control rackets to bounce ball over a net.
Variants: Tennis-Sparse-2T1P-{Discrete,Continuous}
Goal: The agents must not let the ball touch the ground of their own side.
Agents: The environment contains two agents with same Behavior Parameters.
Agent Reward Function:
- +0.0 To the agent whose ground is touched by the ball.
- +1.0 To the other agent.
Episode Terminate Condition:
- The ball touched the ground.
Behavior Parameters:
- Vector Observations:
  - Lidar
- Vector Action space:
  - Discrete: General Player Discrete.
  - Continuous: General Player Continuous.
- Visual Observations:
  - Visual_FP
  - Visual_FP
Reset Parameters: Three
- angle: Angle of the racket from the vertical (Y) axis.
  - Default: 55
  - Recommended Minimum: 35
  - Recommended Maximum: 65
- gravity: Magnitude of gravity
  - Default: 9.81
  - Recommended Minimum: 6
  - Recommended Maximum: 20
- scale: Specifies the scale of the ball in the 3 dimensions (equal across the three dimensions)
  - Default: 1
  - Recommended Minimum: 0.2
  - Recommended Maximum: 5
Benchmark Mean Reward: 2.5
Config
- python train.py -f ./arena-experiments/Benchmark-2T1P-Discrete.yaml
restore
- python train.py -f ./arena-experiments/Benchmark-2T1P-Discrete.yaml --resume

Provide feedback