Api thermal heuristic by Juliette-Gerbaux · Pull Request #238 · AntaresSimulatorTeam/GemsPy

Juliette-Gerbaux · 2026-06-19T07:51:02Z

Process ID

Process:

Description

Impact Analysis

Checklist

Unit tests pass (pytest)
Type checking passes (mypy)
Formatting passes (black, isort)
pyproject.toml version bumped if applicable
AGENTS.md reviewed for impact and updated if needed

Cross-check each heuristic input/output declared in optim-config.yml against the model: the referenced id must exist and have the time-dependence the fast/accurate thermal heuristics expect (e.g. min_up_duration constant, generation_power per-timestep), catching mismatches at load time instead of a runtime crash mid-solve. nb_units_max and cluster_max_generation now accept either form, with the heuristics broadcasting a scalar internally. Also warn (instead of silently truncating) when min_up_duration/ min_down_duration resolve to a non-integer number of timesteps. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

tbittar

We may flag all e2e heuristic test so that we are able to deactivate them easily from the CI at some point (if they get too long)

tbittar · 2026-07-30T11:28:15Z

+            "nb_units_on_opt",
+            "nb_units_max",


To use the same naming conventions as in the models libraries

Suggested change

"nb_units_on_opt",

"nb_units_max",

"num_units_on_opt",

"num_units_max",

tbittar · 2026-07-30T11:28:33Z

+            "min_up_duration",
+            "min_down_duration",
+        },
+        "outputs": {"minimum_nb_units_on"},


Suggested change

"outputs": {"minimum_nb_units_on"},

"outputs": {"minimum_num_units_on"},

tbittar · 2026-07-30T11:40:18Z

    id: str
    model_decomposition: Optional[ModelDecompositionConfig] = None
    out_of_bounds_processing: Optional[OutOfBoundsProcessingConfig] = None
+    heuristic: List[HeuristicConfig] = Field(default_factory=list)


As we may have many of them

Suggested change

heuristic: List[HeuristicConfig] = Field(default_factory=list)

heuristics: List[HeuristicConfig] = Field(default_factory=list)

tbittar · 2026-07-30T11:56:11Z

+                v_cfg.id: v_cfg.location
+                for v_cfg in model_config.model_decomposition.variables
+            }


Suggested change

v_cfg.id: v_cfg.location

for v_cfg in model_config.model_decomposition.variables

}

variable_config.id: variable_config.location

for variable_config in model_config.model_decomposition.variables

}

tbittar · 2026-07-30T16:16:36Z

+    Model on 168 time steps with one thermal generation and one demand on a single node.
+        - Demand is constant to 2000 MW except for the 13th hour for which it is 2050 MW
+        - Thermal generation is characterized with:
+            - P_min = 700 MW
+            - P_max = 1000 MW
+            - Min up time = 3
+            - Min down time = 10
+            - Generation cost = 50€ / MWh
+            - Startup cost = 50
+            - Fixed cost = 1 /h
+            - Number of unit = 3
+        - Unsupplied energy = 1000 €/MWh
+        - Spillage = 0 €/MWh


Expose the system description only once in a comment at the begining of the file

tbittar · 2026-07-30T16:23:26Z

+        "num_units_on",
+        [
+            0.0,
+            1.0,
+            1.0,
+            1.04,
+            1.07,
+            1.11,
+            1.14,
+            1.11,
+            1.08,
+            1.04,
+            1.0,
+            1.0,
+            0.0,
+        ],


I am a bit surprised of this behavior, I would have expected to retrieve the MILP solution of turning on more units (but an integer number of them)

Or maybe this behavior is linked to the addition of ramp constraints ? Is the goal to show the inefficiency of the accurate heuristics in this case ?

tbittar · 2026-07-30T16:26:47Z

+
+def test_accurate_heuristic() -> None:
+    """
+    Check that find_nb_units_accurate produces the same nb_units_on as Antares.


remove Antares reference, just say the goal of the test is to retrieve the expected results stored in the repo

tbittar · 2026-07-30T16:27:27Z

+# This file is part of the Antares project.
+
+# Expected per-timestep outputs for thermal_heuristic_three_clusters tests.
+# Extracted from Antares reference outputs (commit 38c4535).


Remove reference to Antares, just say these are reference results

tbittar · 2026-07-30T16:29:20Z

+    return series[:NUMBER_HOURS, SCENARIO].tolist()
+
+
+def test_accurate_heuristic() -> None:


Move this file to the unittests directory, then what is the added value of having the previous file testing the fast heuristic ?

tbittar · 2026-07-31T17:03:07Z

Fix solution retrieval, window size in fast heuristic bugs (`b96b782`)

Follow-up commit addressing 4 issues surfaced by /code-review on this branch:

1. Root cause: solution reads keyed by an ambiguous merged-variable name
For a model that splits an INTEGER/BINARY variable across relaxed and exact integer_strategy groups, optimization.py registers two real linopy.Variables but merges them into a synthetic xr.concat-ed copy for the general linopy_vars lookup dict. That copy's .name arbitrarily reflects only one of the two real registered names. Anything that read solved values via linopy_model.solution[<name>] (heuristic variable-solution inputs, sequential-mode carry-over extraction, and the simulation table's output/extra-output collectors) would silently get NaN or drop the other group's components.

Fixed by adding OptimizationProblem.get_variable_solution(model_id, var_name), which reads .solution directly off the real per-component-registered variables (_linopy_vars_by_component) and reassembles them — the same mechanism already used for safe bound mutation via get_component_variable. Rewired all affected call sites in heuristic_runner.py, session.py, and simulation_table.py.

2. Heuristics/second solve ran before checking the first solve's status
SimulationSession._run_block applied heuristics and re-solved unconditionally before ever checking whether the first solve reached optimality. Moved the status check to run immediately after the first solve (via a small _check_solved helper), so a failed first solve fails fast instead of running the heuristic step against a non-optimal or missing solution.

3. Fast heuristic crashed on zero minimum durations
find_min_generation_fast computed window_size = max(min_up_duration, min_down_duration), which is 0 for a cluster with no minimum up/down time, and used it as a range() step — raising ValueError: range() arg 3 must not be zero. Clamped to max(..., 1), which degenerates correctly to per-timestep processing.

4. Minor: missing space in the solve-failure error message.

Also updated the FakeProblem test doubles in test_simulation_table_{accessor,export,mock}.py to implement get_variable_solution, matching the new interface.

Verification: full suite green (605 passed, 1 xfailed), mypy clean on all touched files.

Juliette-Gerbaux added 30 commits June 28, 2024 12:22

Add first tests on thermal heuristics

d2b46d2

Implement accurate heuristic on simple problem

24700f4

Add comments

fd3a884

Add fast heuristic and improve accurate

0a6d4df

rename test heuristic

c703040

Implement complex model

9db4b54

Accurate for complex model

83360de

Type errors and useless solver parameters

bde9904

check hourly outputs

4bfd696

fast heuristic for complex case

3427eb4

Test for 2 weeks

c79ba8f

Test for 2 scenarios

dd61b8d

add test for scenarios

5cb88db

Add more details on tests concerning scenarios

9a9ec8d

Add milp test for second case

1be8483

remove milp test

96fa166

Add test accurate for second test case

add82e9

Add fast test for second test case

40282a0

convert fast to optimization problem

5700856

Optimization problem for fast on simple case

dcd2c47

Optimization model fast complex case

ea55b7d

Small corrections

25bddc0

Correct test

db17284

Optimization problem heuristic fast second complex test case

6cf6611

Equivalent solution for fast heuristic optimization

3d9cb5b

Api for fast and accurate models

57e5f7c

Api for accurate heuristic model

e828f65

Move models to library

78a4712

Move some functions from tests to src

0b10902

Remove all functions from simple test

912f083

Juliette-Gerbaux and others added 23 commits June 17, 2026 10:31

Merge branch 'main' into api_thermal_heuristic

38c4535

Move files

a8271fb

Prepare test_one_cluster

2015411

Update tests

81fb135

Update test with bc

e50a82f

Update test with different scenarios and xpress settings

05e3791

Remove test with different scenarios and xpress settings

2622caa

Remove day ahead test

cb853e4

Implement fast heuristic as an algorithm

85bf988

Implement accurate heuristic

46b3805

integer-strategy in system and heuristic in optim-config

54b9fed

Implement interger-strategy in variables building

11652f3

heuristic id in parsing

19a0128

Add heuristic validation and update optim-config documentation

7eb031c

Workflow with 2 iterations and thermal heuristic

f81b7a9

Remove obsolete code

bd43257

Fix ci

ff9533d

Test with ramp

db5ef38

Validate optim-config automatically in SimulationSession

e2f88ec

Merge branch 'main' into api_thermal_heuristic

15456c0

Fix ci

decc740

Refactor e2e tests

3320444

Juliette-Gerbaux marked this pull request as ready for review July 24, 2026 16:53

Juliette-Gerbaux requested a review from tbittar July 24, 2026 16:54

Formatting

d45ce1a

tbittar reviewed Jul 30, 2026

View reviewed changes

tbittar added 2 commits July 31, 2026 16:19

Refactoring

bcecb33

Fix solution retrieval, window size in fast heuristic bugs

b96b782

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Api thermal heuristic - #238

Api thermal heuristic#238
Juliette-Gerbaux wants to merge 120 commits into
mainfrom
api_thermal_heuristic

Juliette-Gerbaux commented Jun 19, 2026

Uh oh!

tbittar left a comment •

edited

Loading

Uh oh!

tbittar Jul 30, 2026

Uh oh!

tbittar Jul 30, 2026

Uh oh!

tbittar Jul 30, 2026

Uh oh!

Uh oh!

tbittar Jul 30, 2026

Uh oh!

tbittar Jul 30, 2026

Uh oh!

tbittar Jul 30, 2026

Uh oh!

tbittar Jul 30, 2026

Uh oh!

tbittar Jul 30, 2026

Uh oh!

tbittar Jul 30, 2026

Uh oh!

tbittar commented Jul 31, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

	"outputs": {"minimum_nb_units_on"},
	"outputs": {"minimum_num_units_on"},

	heuristic: List[HeuristicConfig] = Field(default_factory=list)
	heuristics: List[HeuristicConfig] = Field(default_factory=list)

		return series[:NUMBER_HOURS, SCENARIO].tolist()


		def test_accurate_heuristic() -> None:

Uh oh!

Conversation

Juliette-Gerbaux commented Jun 19, 2026

Process ID

Description

Impact Analysis

Checklist

Uh oh!

tbittar left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

tbittar commented Jul 31, 2026

Fix solution retrieval, window size in fast heuristic bugs (b96b782)

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

tbittar left a comment •

edited

Loading

Fix solution retrieval, window size in fast heuristic bugs (`b96b782`)