passagemath-glpk

10.8.4 verified Mon Apr 27 auth: no python

passagemath-glpk provides a linear and mixed integer linear optimization backend using GLPK (GNU Linear Programming Kit) for the passagemath library. It enables the construction and solution of linear programming (LP) and mixed-integer linear programming (MILP) problems. Version 10.8.4 supports Python >=3.11,<3.15. The package is part of the passagemath ecosystem, which reimplements functionality previously provided by SageMath.

pip install passagemath-glpk

Common errors

error ImportError: cannot import name 'GLPKBackend' from 'passagemath.glpk.backend' ↓

cause GLPKBackend not available if passagemath-glpk not installed or version mismatch.

fix

Run 'pip install passagemath-glpk' and ensure passagemath core is also installed.

error ValueError: solver 'GLPK' is not available ↓

cause The GLPK solver is not installed or not recognized by passagemath.

fix

Install passagemath-glpk: 'pip install passagemath-glpk'.

error TypeError: __init__() got an unexpected keyword argument 'solver' ↓

cause Using old SageMath-style instantiation without solver argument.

fix

Use 'MixedIntegerLinearProgram(solver='GLPK')' instead of 'MixedIntegerLinearProgram()'.

Warnings

breaking GLPKBackend import path changed from sage.numerical.backends.glpk_backend to passagemath.glpk.backend. ↓

fix Change import to 'from passagemath.glpk.backend import GLPKBackend'.

breaking MixedIntegerLinearProgram now requires 'solver' parameter; default may not be GLPK. In SageMath, default was GLPK. ↓

fix Explicitly pass solver='GLPK' when creating MixedIntegerLinearProgram.

gotcha GLPK backend may not be thread-safe; avoid concurrent solve calls. ↓

fix Serialize solve calls or use multiprocessing with separate processes.

Imports

GLPKBackend

wrong

from sage.numerical.backends.glpk_backend import GLPKBackend

correct

from passagemath.glpk.backend import GLPKBackend

Old SageMath import path won't work; passagemath uses 'passagemath.glpk.backend'.

MixedIntegerLinearProgram

wrong

from sage.numerical.mip import MixedIntegerLinearProgram

correct

from passagemath.numerical.mip import MixedIntegerLinearProgram

SageMath import path replaced; use passagemath.numerical.mip.

GLPK

wrong

from sage.numerical.backends.glpk_backend import GLPKBackend

correct

from passagemath.glpk.backend import GLPKBackend

Correct class is GLPKBackend; common mistake to use GLPK directly.

Quickstart

Quickstart example: solve a simple MILP using GLPK backend.

from passagemath.numerical.mip import MixedIntegerLinearProgram

# Create MILP instance with GLPK backend
p = MixedIntegerLinearProgram(solver='GLPK')

# Define variables
x = p['x']
y = p['y']

# Add constraints
p.add_constraint(x + 2*y <= 4)
p.add_constraint(2*x + y <= 6)

# Set objective
p.set_objective(3*x + 4*y)

# Solve
p.solve()

# Get solution
print(f"x = {p.get_values(x)}")
print(f"y = {p.get_values(y)}")