Downloading a MIPLIB Instance

The OMMX repository provides mixed-integer programming benchmark instances from MIPLIB 2017 in OMMX Artifact format.

Note

More details: The MIPLIB 2017 instances in OMMX Artifact format are hosted in the GitHub Container Registry for the OMMX repository (link).

Please see this page for information on GitHub Container Registry.

You can easily download these instances with the OMMX SDK, then directly use them as inputs to OMMX Adapters. For example, to solve the air05 instance from MIPLIB 2017 (reference) with PySCIPOpt, you can:

1. Download the air05 instance with dataset.miplib2017 from the OMMX Python SDK.

2. Solve with PySCIPOpt via the OMMX PySCIPOpt Adapter.

Here is a sample Python code:

# OMMX Python SDK
from ommx import dataset
# OMMX PySCIPOpt Adapter
from ommx_pyscipopt_adapter import OMMXPySCIPOptAdapter

# Step 1: Download the air05 instance from MIPLIB 2017
instance = dataset.miplib2017("air05")

# Step 2: Solve with PySCIPOpt via the OMMX PySCIPOpt Adapter
solution = OMMXPySCIPOptAdapter.solve(instance)

This functionality makes it easy to run benchmark tests on multiple OMMX-compatible solvers using the same MIPLIB instances.

Note about Annotations with the Instance

The downloaded instance includes various annotations accessible via the annotations property:

import pandas as pd
# Display annotations in tabular form using pandas
pd.DataFrame.from_dict(instance.annotations, orient="index", columns=["Value"]).sort_index()

	Value
org.ommx.miplib.binaries	7195
org.ommx.miplib.continuous	0
org.ommx.miplib.group	air
org.ommx.miplib.integers	0
org.ommx.miplib.non_zero	52121
org.ommx.miplib.objective	26374
org.ommx.miplib.status	easy
org.ommx.miplib.tags	benchmark,binary,benchmark_suitable,set_partit...
org.ommx.miplib.url	https://miplib.zib.de/instance_details_air05.html
org.ommx.v1.instance.authors	G. Astfalk
org.ommx.v1.instance.constraints	426
org.ommx.v1.instance.created	2024-11-11T20:38:59.242221+09:00
org.ommx.v1.instance.dataset	MIPLIB2017
org.ommx.v1.instance.license	CC-BY-SA-4.0
org.ommx.v1.instance.title	air05
org.ommx.v1.instance.variables	7195

These instances have both dataset-level annotations and dataset-specific annotations.

There are seven dataset-wide annotations with dedicated properties:

Annotation	Property	Description
org.ommx.v1.instance.authors	authors	The authors of the instance
org.ommx.v1.instance.constraints	num_constraints	The number of constraint conditions in the instance
org.ommx.v1.instance.created	created	The date of the instance was saved as an OMMX Artifact
org.ommx.v1.instance.dataset	dataset	The name of the dataset to which this instance belongs
org.ommx.v1.instance.license	license	The license of this dataset
org.ommx.v1.instance.title	title	The name of the instance
org.ommx.v1.instance.variables	num_variables	The total number of decision variables in the instance

MIPLIB-specific annotations are prefixed with org.ommx.miplib.*.

For example, the optimal objective of the air05 instance is 26374, which you can check with the key org.ommx.miplib.objective:

# Note that the values of annotations are all strings (str)!
assert instance.annotations["org.ommx.miplib.objective"] == "26374"

Thus, we can verify that the optimization result from the OMMX PySCIPOpt Adapter matches the expected optimal value.

import numpy as np

best = float(instance.annotations["org.ommx.miplib.objective"])
assert np.isclose(solution.objective, best)