---
license: mit
library_name: joblib
tags:
- batteryswapai-2026
- predictive-maintenance
- survival-analysis
- scheduling
---

# BatterySwapAI 2026 — MnesisLab

Causal battery-swap planning: hierarchical Wiener first-passage reranking over a
degradation model, with a cost-aware capacity and routing policy.

## Artifact

`submission_artifacts/weekly99_planner.joblib`, loaded by `script.py`.

| setting | value |
|---|---|
| FPT rank residual weight | 0.15 |
| capacity lookback | 42 days |
| emergency operational scale | 0.5 |
| capacity weekly limit fraction | 0.99 |

## Contract

- Each scenario uses only readings with `end_time <= scenario.start_time`.
- EOL is reconstructed as the evaluator defines it: strict `10 < T < 30`, daily median,
  days with fewer than five readings masked, seven-calendar-day rolling median with
  `min_periods=3`, first smoothed voltage `<= 2.40 V`. This matches all 82 observed train
  EOL devices; censored devices remain censored.
- Runtime: `batteryswap_public==0.3.4`, CPU only, no network, every live battery emitted
  once with a valid plan date, 19,890 rows on the train split.

MIT licensed (`LICENSE`). Third-party notices in `THIRD_PARTY_LICENSES.md`.