| """Inspect compatible improvement in the supplied finite synthetic arena.""" |
| from pathlib import Path |
| import sys |
| sys.path.insert(0, str(Path(__file__).resolve().parents[1] / 'src')) |
| import numpy as np |
| from infinite_players.arena import make_arena |
| from infinite_players.fidelity import policy_iteration, semantic_law, signature_residual |
|
|
|
|
| def main(): |
| mdp, anchor, targets = make_arena(seed=1, theta=1) |
| descendant, value, audit = policy_iteration(mdp, targets, kind='path') |
| source_value, _ = mdp.evaluate(anchor) |
| path_tv = np.abs(semantic_law(mdp, anchor, 6) - semantic_law(mdp, descendant, 6)).sum()/2 |
| print(f'Normalized source value: {(1-mdp.gamma)*source_value.mean():.6f}') |
| print(f'Normalized descendant value: {(1-mdp.gamma)*value.mean():.6f}') |
| print(f'Six-step semantic TV: {path_tv:.3e}') |
| print(f'Maximum signature residual: {signature_residual(mdp, descendant, targets):.3e}') |
| print(f'Policy iteration steps: {audit["iterations"]}') |
| print('Scope: one synthetic known-model arena; no human or commercial-game evaluation.') |
|
|
|
|
| if __name__ == '__main__': |
| main() |
|
|
