| from __future__ import annotations |
|
|
| import numpy as np |
| import pandas as pd |
|
|
|
|
| def max_drawdown(values: list[float] | np.ndarray) -> float: |
| series = np.asarray(values, dtype=np.float64) |
| if series.size == 0: |
| return 0.0 |
| peaks = np.maximum.accumulate(series) |
| return float(np.max(1.0 - series / np.maximum(peaks, 1e-12))) |
|
|
|
|
| def action_jitter(actions: list[list[float]] | np.ndarray) -> float: |
| arr = np.asarray(actions, dtype=np.float64) |
| if arr.ndim != 2 or len(arr) < 2: |
| return 0.0 |
| return float(np.mean(np.linalg.norm(np.diff(arr, axis=0), ord=1, axis=1))) |
|
|
|
|
| def spatial_fairness_index(city_rewards: dict[str, float]) -> float: |
| values = np.asarray(list(city_rewards.values()), dtype=np.float64) |
| if values.size == 0: |
| return 0.0 |
| mean_abs = max(float(np.mean(np.abs(values))), 1e-9) |
| return float(np.var(values) / mean_abs) |
|
|
|
|
| def summarize_episode(rows: list[dict]) -> dict[str, float]: |
| if not rows: |
| return {} |
| frame = pd.DataFrame(rows) |
| city_rewards: dict[str, float] = {} |
| for row in rows: |
| for city, value in row.get("city_rewards", {}).items(): |
| city_rewards[city] = city_rewards.get(city, 0.0) + float(value) |
|
|
| return { |
| "cumulative_reward": float(frame["reward"].sum()), |
| "episode_volume": float(frame["matched_energy_MWh"].sum()), |
| "physics_violation_rate": float(frame["physics_violation_rate"].mean()), |
| "max_drawdown": max_drawdown(frame["liquidity"].to_numpy()), |
| "max_token_drawdown": max_drawdown(frame["token_price"].to_numpy()), |
| "action_jitter": action_jitter(frame[["reward_ratio", "liquidity_ratio", "burn_rate"]].to_numpy()), |
| "mean_slippage": float(frame["slippage"].mean()), |
| "spatial_fairness_index": spatial_fairness_index(city_rewards), |
| "artificial_liquidity_MWh": float(frame["artificial_liquidity_MWh"].sum()), |
| "final_liquidity": float(frame["liquidity"].iloc[-1]), |
| "final_token_price": float(frame["token_price"].iloc[-1]), |
| } |
|
|
