SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2605.06557 · 2026

Coordination Matters: Evaluation of Cooperative Multi-Agent Reinforcement Learning

Afsaneh Doryab, Maria Cardei, Matthew Landers

arXiv · PDF · Open in the Atlas

Code that ran

8 of 11 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 3 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 5 failures of our instrument; 3 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2605.06557.json · badge: /api/ran/2605.06557.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
mariacardei/coordination_aware_MARL canonical (classifier label)
classifier basis: census
8 ran · 3 no recorded run
FunctionStatusWhere it lives
compute_completed Ran
ran, honoured the contract we drafted
mariacardei/coordination_aware_MARL/src/stat_env/stat_model.py
code served (permissive licence) · get_code("d3fa5e5acb47f542")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

class Task:
    def __init__(self, completed):
        self.completed = completed

class MockModel:
    def __init__(self, sorted_tasks):
        self.sorted_tasks = sorted_tasks

# Create a few task instances with different completion statuses
task1 = Task(completed=True)
task2 = Task(completed=False)
task3 = Task(completed=True)
task4 = Task(completed=True)

# Create the model with sorted_tasks attribute
model = MockModel(sorted_tasks=[task1, task2, task3, task4])

ARGS = {
    "model": model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_env_from_config Ran
ran, honoured the contract we drafted
mariacardei/coordination_aware_MARL/src/fdqn/eval_fdqn.py
code served (permissive licence) · get_code("f44e5d5a2251a9c8")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family int; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re

def extract_env_from_config(config: str):
    """
    config expected like: 3R_5V_5x5
    """
    m = re.match('^(\\d+)R_(\\d+)V_(\\d+)x(\\d+)$', config)
    if not m:
        return None
    return (int(m.group(1)), int(m.group(2)), int(m.group(3)), int(m.group(4)))

ARGS = {
    "config": "3R_5V_5x5"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family int — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_env_from_filename Ran
ran, honoured the contract we drafted
mariacardei/coordination_aware_MARL/src/dqn/eval_dqn.py
code served (permissive licence) · get_code("40bd9356590d7058")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family int; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re

def extract_env_from_filename(filename: str):
    m = re.match('(\\d+)R_(\\d+)V_(\\d+)x(\\d+)\\.pth$', filename)
    if not m:
        return None
    return (int(m.group(1)), int(m.group(2)), int(m.group(3)), int(m.group(4)))

# Valid filename that matches the pattern: (digits)R_(digits)V_(digits)x(digits).pth
ARGS = {
    "filename": "42R_17V_5x3.pth"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family int — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_train_seed_from_path Ran
ran; our contract draft was wrong, not the code
mariacardei/coordination_aware_MARL/src/dqn/eval_dqn.py
code served (permissive licence) · get_code("e6aeb129ca653517")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re

def extract_train_seed_from_path(model_path: str) -> str:
    m = re.search('(seed\\d+)', model_path)
    return m.group(1) if m else 'seedUNK'

ARGS = {
    'model_path': '/path/to/model_seed42_checkpoint.pt'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
find_run_dir Ran
ran; our contract draft was wrong, not the code
mariacardei/coordination_aware_MARL/src/fdqn/eval_fdqn.py
code served (permissive licence) · get_code("638b62b4310416b3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import tempfile
import shutil

# Create a temporary directory structure that mimics the expected layout
temp_root = tempfile.mkdtemp()

# Create the expected directory structure
fdqn_root = temp_root
training_experiments_dir = os.path.join(fdqn_root, 'training_experiments')
os.makedirs(training_experiments_dir)

config_name = 'test_config'
exp_root = os.path.join(training_experiments_dir, f'FDQN_{config_name}')
os.makedirs(exp_root)

# Create a seed directory with the expected naming pattern
train_seed = 42
seed_dir_name = f'seed{train_seed}_run_0'
seed_dir_path = os.path.join(exp_root, seed_dir_name)
os.makedirs(seed_dir_path)

ARGS = {
    'fdqn_root': fdqn_root,
    'config': config_name,
    'train_seed': train_seed,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
find_trained_model Ran
ran; our contract draft was wrong, not the code
mariacardei/coordination_aware_MARL/src/dqn/eval_dqn.py
code served (permissive licence) · get_code("e905b5905d875150")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import tempfile
import shutil

# Create a temporary directory structure that mimics the expected layout
temp_dir = tempfile.mkdtemp()

dqn_root = temp_dir
config = "CartPole-v1"
train_seed = 42

# Create the expected directory structure
exp_root = os.path.join(dqn_root, 'training_experiments', f'DQN_{config}')
os.makedirs(exp_root, exist_ok=True)

# Create a seed directory
seed_dir = os.path.join(exp_root, f'seed{train_seed}_run0')
os.makedirs(seed_dir, exist_ok=True)

# Create the saved_models directory and a dummy model file
saved_models_dir = os.path.join(seed_dir, 'saved_models')
os.makedirs(saved_models_dir, exist_ok=True)

model_path = os.path.join(saved_models_dir, f'{config}.pth')
# Create a dummy model file
with open(model_path, 'w') as f:
    f.write('dummy_model')

ARGS = {
    'dqn_root': dqn_root,
    'config': config,
    'train_seed': train_seed,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_distance Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
mariacardei/coordination_aware_MARL/src/stat_env/stat_model.py
code served (permissive licence) · get_code("560f29de1e064604")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import math

def get_distance(pos1, pos2):
    """Return Euclidean distance between two positions."""
    x_1, y_1 = pos1
    x_2, y_2 = pos2
    return math.sqrt((x_2 - x_1) ** 2 + (y_2 - y_1) ** 2)


# Construct valid arguments
pos1 = (0.0, 0.0)
pos2 = (3.0, 4.0)

ARGS = {
    "pos1": pos1,
    "pos2": pos2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
infer_metrics_from_csv Ran
ran; our contract draft was wrong, not the code
mariacardei/coordination_aware_MARL/src/dqn/run_all_evals.py
code served (permissive licence) · get_code("15a58d2465e21e51")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import pandas as pd

# Create a simple DataFrame with columns that include ID_COLS and metrics
df = pd.DataFrame({
    'config': ['config_a', 'config_b'],
    'train_seed': [42, 43],
    'eval_seed': [100, 101],
    'episode_idx': [1, 2],
    'accuracy': [0.95, 0.97],
    'loss': [0.05, 0.03],
    'reward': [10.5, 11.2],
})

ARGS = {
    'df': df
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
calc_move Not yet run
unverified (harvested, no recorded run)
mariacardei/coordination_aware_MARL/src/stat_env/stat_model.py
code served (permissive licence) · get_code("a698f6a10eefce1e")
How it ran

no invocation recorded.

make_dqn_agent Not yet run
unverified (harvested, no recorded run)
mariacardei/coordination_aware_MARL/src/dqn/train_dqn.py
code served (permissive licence) · get_code("cf8c34d90dbdd3de")
How it ran

no invocation recorded.

make_fdqn_agent Not yet run
unverified (harvested, no recorded run)
mariacardei/coordination_aware_MARL/src/fdqn/train_fdqn.py
code served (permissive licence) · get_code("6fd79f8e5f3fee61")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 11 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Cooperative multi-agent reinforcement learning (MARL) benchmarks commonly emphasize aggregate outcomes such as return, success rate, or completion time. While essential, these metrics often fail to reveal how agents coordinate, particularly in settings where agents, tasks, and joint assignment choices scale combinatorially. We propose a coordination-aware evaluation perspective that supplements return with process-level diagnostics. We instantiate this perspective using STAT, a controlled commitment-constrained spatial task-allocation testbed that systematically varies agents, tasks, and environment size while holding observation access and task rules fixed. We evaluate six representative value-based MARL methods across varying levels of centralization. Our results show that similar return trends can reflect distinct coordination mechanisms, including differences in redundant assignment, assignment diversity, and task-completion efficiency. We find that in commitment-constrained task allocation, performance under scale is shaped not only by nominal action-space size, but also by assignment pressure, sparse decision opportunities, and redundant choices among interdependent agents. Our findings motivate coordination-aware evaluation as a necessary complement to return-based benchmarking for cooperative MARL. 2

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2605.06557")
get_code_for_paper("2605.06557")
have("2605.06557")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2605.06557.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2605.06557.svg)](https://syntology.ai/paper/2605.06557)

Connect an agent — have() is free.