seed: 20260511
episode_steps: 24
data_dir: data/datasets
output_dir: outputs/runs
figures_dir: figures
action_mode: continuous
discrete_levels: 5
no_physics_penalty: false

market:
  initial_liquidity: 1.0
  initial_token_price: 1.0
  min_reward_ratio: 0.05
  max_reward_ratio: 0.65
  min_liquidity_ratio: 0.20
  max_liquidity_ratio: 0.95
  max_burn_rate: 0.20
  max_total_allocation: 0.98
  static_reward_ratio: 0.25
  static_liquidity_ratio: 0.75
  static_burn_rate: 0.02

reward:
  drawdown_penalty: 1.50
  action_jitter_penalty: 0.25
  unmet_demand_penalty: 0.40
  physics_penalty: 2.00
  fairness_penalty: 0.20

training:
  timesteps: 10000
  n_envs: 1
  learning_rate: 0.0003
  batch_size: 64
  gamma: 0.98

evaluation:
  episodes: 5
  policies:
    - static
    - random
    - myopic