Skip to main content13 runs
1 runs
4 runs
Experiment Registry
All experiment runs are immutable records. Once created, run records are never updated — corrections create new versioned runs. Metrics displayed here are read from frozen_metrics_snapshot, not recomputed.
18
Total Runs
3
Groups
15
Algorithms
16
Completed
Compare runs side-by-sideSelect any two or more runs from the same experiment group.
Open Compare Engine →Multi-Algorithm Benchmark (13 algos)
rl_multialgo_20251127| Run ID | Algorithm | Dataset | Cost ↓% | Convergence | Stability | Artifacts | Determinism | Status | |
|---|---|---|---|---|---|---|---|---|---|
| rl_a3c_20251127_162240_r1 | A3C | TEIAS v2025.1 | −97.82% | 4724 | 72 | 2 | Seeded | ✓ Completed | View → |
| rl_marl_20251128_003215_r1 | MARL | TEIAS v2025.1 | −97.82% | 5123 | 85 | 2 | Seeded | ✓ Completed | View → |
| rl_pso_20251127_210312_r1 | PSO | TEIAS v2025.1 | −97.77% | 4842 | 82 | 1 | Seeded | ✓ Completed | View → |
| rl_dqn_20251127_191023_r1 | DQN | TEIAS v2025.1 | −97.69% | 4913 | 76 | 2 | Seeded | ✓ Completed | View → |
| rl_sac_20251127_131802_r1 | SAC | TEIAS v2025.1 | −97.58% | 5271 | 87 | 3 | Seeded | ✓ Completed | View → |
| rl_ppo_20251127_114523_r1 | PPO | TEIAS v2025.1 | −97.34% | 4924 | 98 | 4 | Seeded | ✓ Completed | View → |
| rl_ga_20251127_200145_r1 | GA | TEIAS v2025.1 | −97.28% | 4497 | 88 | 1 | Seeded | ✓ Completed | View → |
| rl_de_20251127_220451_r1 | DE | TEIAS v2025.1 | −97.22% | 4380 | 87 | 1 | Seeded | ✓ Completed | View → |
| rl_fuzzyql_20251127_231015_r1 | FuzzyQL | TEIAS v2025.1 | −97.21% | 4448 | 94 | 1 | Seeded | ✓ Completed | View → |
| rl_neurofuzzy_20251127_235022_r1 | NF | TEIAS v2025.1 | −97.21% | 4430 | 96 | 1 | Seeded | ✓ Completed | View → |
| rl_pnn_20251128_013845_r1 | PNN | TEIAS v2025.1 | −97.21% | 4495 | 91 | 1 | Seeded | ✓ Completed | View → |
| rl_maml_20251128_010430_r1 | MAML | TEIAS v2025.1 | −97.19% | 4478 | 92 | 1 | Seeded | ✓ Completed | View → |
| rl_td3_20251127_145631_r1 | TD3 | TEIAS v2025.1 | −96.32% | 4100 | 85 | 2 | Seeded | ✓ Completed | View → |
ENSM/HNSM Forecast Study
forecast_ensm_20251017| Run ID | Algorithm | Dataset | Cost ↓% | Convergence | Stability | Artifacts | Determinism | Status | |
|---|---|---|---|---|---|---|---|---|---|
| forecast_ensm_v40_20251017 | ENSM | Turkey Energy v2024 | R²=0.9799 | 99 | 99 | 6 | Deterministic | ✓ Completed | View → |
Validation & Stress Tests
validation_runs| Run ID | Algorithm | Dataset | Cost ↓% | Convergence | Stability | Artifacts | Determinism | Status | |
|---|---|---|---|---|---|---|---|---|---|
| rl_dqn_discrete_5lvl_20251127 | DQN | TEIAS v2025.1 | −97.69% | 4913 | 76 | 1 | Seeded | ✓ Completed | View → |
| rl_a3c_highre_20251103 | A3C | TEIAS v2025.1 | −91.2% | 3400 | 48 | 1 | Seeded | ⚠ Partial | View → |
| rl_ppo_instability_test_01 | PPO | TEIAS v2025.1 | −84.5% | 2100 | 32 | 1 | Seeded | ⚠ Partial | View → |
| forecast_lstm_v2_validation_20250815 | LSTM | EPİAŞ v2025.1 | R²=0.8932 | 89 | 99 | 1 | Seeded | ✓ Completed | View → |
NON-NEGOTIABLE: Immutability — experiment run records are INSERT-only. Re-importing the same source file is idempotent (no duplicate). Corrections create a new run with a supersedes_run_id link. This rule is enforced at the database level via a PostgreSQL BEFORE UPDATE trigger. See the Reproducibility hub for details.