challenge_ai

3 Commits 1 Branch 0 Tags

3 Commits

Author	SHA1	Message	Date
mattandClaude Sonnet 5	afa02864a7	Fix test suite hangs caused by degenerate MILP fixtures CI / test (push) Successful in 12s Details Two tests were spinning up real CBC solves over 96-half-hour windows with long runs of exactly-repeated prices (e.g. [0.0] * 40). That gives the LP relaxation a huge set of economically indistinguishable ways to spread a trade, which the MILP fallback's branch-and-bound then wastes enormous effort disambiguating (47k+ nodes without closing the gap, confirmed by running CBC verbosely). Real Attachment 2 data has no such flat runs and solves in ~0.1s/window; a synthetic sine wiggle wasn't enough either, since neighbouring half-hours stayed too similar. Fixes, matched to what each test actually needs: - The 9 validator tests only need some structurally valid schedule to mutate; they were deriving it by running the real optimiser once per test. Replaced with make_valid_schedule(), built directly from the model's own energy-balance formulas -- no solver involved, and it's now a true unit test of validate_schedule() in isolation. - The rolling-horizon carry-forward test was exercising the mechanism at full production scale (48h window / 24h commit) when a 2h/1h window proves the same boundary-carrying behaviour with a trivial MILP, regardless of price structure. - Fixed a genuine tie in test_optimum_uses_whichever_market_pays_more: two equal-price hours with just enough stored energy for one meant either market was a valid optimum. Sized the charge phase so delivery must split across both hours, pinning a unique answer. Full suite: 38 passed in ~1.3s (previously hung indefinitely on CI and locally within seconds of the same wall-clock variance CBC shows on degenerate MIPs). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>	2026-09-24 16:10:55 +01:00
mattandClaude Sonnet 5	298724a9d3	Add battery dispatch implementation and Gitea CI pipeline CI / test (push) Failing after 45m4s Details Implements the package from PLAN.md: config/data loading, the LP-first MILP-fallback rolling-horizon optimiser, battery degradation tracking, independent schedule validation, metrics, plots and the CLI, plus the matching test suite. Adds a Gitea Actions workflow (lint + tests) that posts a pass/fail notification to ntfy on every run, and a .gitignore for build/cache artefacts that had been tracked by mistake. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>	2026-09-24 13:32:06 +01:00
mattandClaude Opus 5	54e877db9b	Add exercise documents and battery dispatch implementation plan Commits the three source documents for the battery trading exercise (task brief, battery specification, and three years of half-hourly Market 1 and hourly Market 2 prices) together with PLAN.md, the implementation plan for the dispatch model. The plan covers a rolling-horizon MILP on the half-hourly grid with a 48-hour lookahead and 24-hour commitment, solved with PuLP/CBC, plus degradation and cycle-life reporting, an independent schedule validator, charts and tests. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>	2026-09-24 12:47:12 +01:00

Author

SHA1

Message

Date

mattandClaude Sonnet 5

afa02864a7

Fix test suite hangs caused by degenerate MILP fixtures

CI / test (push) Successful in 12s

Details

Two tests were spinning up real CBC solves over 96-half-hour windows
with long runs of exactly-repeated prices (e.g. [0.0] * 40). That gives
the LP relaxation a huge set of economically indistinguishable ways to
spread a trade, which the MILP fallback's branch-and-bound then wastes
enormous effort disambiguating (47k+ nodes without closing the gap,
confirmed by running CBC verbosely). Real Attachment 2 data has no such
flat runs and solves in ~0.1s/window; a synthetic sine wiggle wasn't
enough either, since neighbouring half-hours stayed too similar.

Fixes, matched to what each test actually needs:
- The 9 validator tests only need *some* structurally valid schedule to
  mutate; they were deriving it by running the real optimiser once per
  test. Replaced with make_valid_schedule(), built directly from the
  model's own energy-balance formulas -- no solver involved, and it's
  now a true unit test of validate_schedule() in isolation.
- The rolling-horizon carry-forward test was exercising the mechanism
  at full production scale (48h window / 24h commit) when a 2h/1h
  window proves the same boundary-carrying behaviour with a trivial
  MILP, regardless of price structure.
- Fixed a genuine tie in test_optimum_uses_whichever_market_pays_more:
  two equal-price hours with just enough stored energy for one meant
  either market was a valid optimum. Sized the charge phase so delivery
  must split across both hours, pinning a unique answer.

Full suite: 38 passed in ~1.3s (previously hung indefinitely on CI and
locally within seconds of the same wall-clock variance CBC shows on
degenerate MIPs).

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

2026-09-24 16:10:55 +01:00

mattandClaude Sonnet 5

298724a9d3

Add battery dispatch implementation and Gitea CI pipeline

CI / test (push) Failing after 45m4s

Details

Implements the package from PLAN.md: config/data loading, the LP-first
MILP-fallback rolling-horizon optimiser, battery degradation tracking,
independent schedule validation, metrics, plots and the CLI, plus the
matching test suite. Adds a Gitea Actions workflow (lint + tests) that
posts a pass/fail notification to ntfy on every run, and a .gitignore
for build/cache artefacts that had been tracked by mistake.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

2026-09-24 13:32:06 +01:00

mattandClaude Opus 5

54e877db9b

Add exercise documents and battery dispatch implementation plan

Commits the three source documents for the battery trading exercise
(task brief, battery specification, and three years of half-hourly
Market 1 and hourly Market 2 prices) together with PLAN.md, the
implementation plan for the dispatch model.

The plan covers a rolling-horizon MILP on the half-hourly grid with a
48-hour lookahead and 24-hour commitment, solved with PuLP/CBC, plus
degradation and cycle-life reporting, an independent schedule
validator, charts and tests.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

2026-09-24 12:47:12 +01:00