PyGLPK

0.4.8 · active · verified Thu Apr 16

PyGLPK (Python GLPK) is a Python module that provides a comprehensive interface to the GNU Linear Programming Kit (GLPK). It allows Python users to model and solve linear programming (LP), mixed integer programming (MIP), and other related optimization problems. The current version is 0.4.8. It has a moderate release cadence, largely tied to maintenance and updates related to the underlying GLPK C library.

Common errors

```
ModuleNotFoundError: No module named 'glpk'
```
cause The PyGLPK Python package has not been installed or is not accessible in the current Python environment.

fix Run `pip install glpk` to install the PyGLPK package.
```
ImportError: libglpk.so.40: cannot open shared object file: No such file or directory (or similar for .dylib on macOS, .dll on Windows)
```
cause The underlying GLPK C library is not found by PyGLPK, likely because it is not installed or not in the system's dynamic link library search path.

fix Install the GLPK C library for your operating system (e.g., `sudo apt-get install libglpk-dev` on Debian/Ubuntu, `brew install glpk` on macOS). Ensure it's correctly linked or added to your system's PATH/LD_LIBRARY_PATH environment variable.
```
glpk.GLPKError: problem has no solution
```
cause The optimization problem is infeasible (no solution satisfies all constraints) or unbounded (the objective function can be improved indefinitely without violating constraints).

fix Review your constraints and objective function for logical errors or contradictions. Use `lp.write_lp("problem.lp")` to export the problem to an LP file and inspect it with a dedicated LP viewer or GLPK's own `glpsol` tool for detailed analysis.

Warnings

gotcha PyGLPK requires the GLPK C library to be installed on your system BEFORE installing the Python package. The Python package is a wrapper; it does not install the underlying C library.
Fix: Install the GLPK C library for your operating system. For Debian/Ubuntu: `sudo apt-get install libglpk-dev`. For macOS (Homebrew): `brew install glpk`. For Windows, you typically need to build from source or use a pre-compiled binary and ensure it's in your system's PATH.
gotcha Be mindful of variable types (continuous, integer, binary) and bounds. Incorrectly defined variables can lead to infeasible or unbounded solutions, or incorrect solver behavior.
Fix: Always explicitly set `kind` (C, I, B for continuous, integer, binary respectively) and `lb` (lower bound) / `ub` (upper bound) for variables. Use `N` from `glpk` for infinity.
gotcha The GLPK solver's status codes (e.g., `lp.status`) are integers. For clear and robust interpretation, always compare them against constants from the `glpk.GLPK` module (e.g., `GLPK.LP_OPT`, `GLPK.LP_FEAS`).
Fix: Use `if lp.status == GLPK.LP_OPT:` instead of `if lp.status == 5:` (or other magic numbers). Refer to GLPK documentation for all status constants.

Install

pip install glpk Install PyGLPK

Imports

GLPK
```
from glpk import GLPK
```
LP
```
from glpk import LP
```
Var
```
from glpk import Var
```
Term
```
from glpk import Term
```
N
```
from glpk import N
```
Represents infinity for variable bounds.
C
```
from glpk import C
```
Represents a continuous variable kind.
R
```
from glpk import R
```
Represents row/constraint.

Quickstart

This quickstart defines and solves a simple linear programming problem with two continuous variables and three constraints. It demonstrates variable definition, objective function setup, constraint creation, solving the problem, and interpreting the results.

from glpk import LP, Var, N, C, GLPK

# Create a new problem
lp = LP()

# Define variables
x1 = lp.add_var(name='x1', lb=0.0, ub=N, kind=C) # lb=0, ub=infinity, continuous
x2 = lp.add_var(name='x2', lb=0.0, ub=N, kind=C)

# Set objective function (maximize 10*x1 + 6*x2)
lp.obj.dir = LP.MAX
lp.obj += 10 * x1 + 6 * x2

# Add constraints
# Constraint 1: x1 + x2 <= 100
c1 = lp.add_row(name='c1')
c1.add_term(x1, 1.0)
c1.add_term(x2, 1.0)
c1.upper_bound = 100.0

# Constraint 2: 7*x1 + 3*x2 <= 350
c2 = lp.add_row(name='c2')
c2.add_term(x1, 7.0)
c2.add_term(x2, 3.0)
c2.upper_bound = 350.0

# Constraint 3: 3*x1 + 5*x2 <= 150
c3 = lp.add_row(name='c3')
c3.add_term(x1, 3.0)
c3.add_term(x2, 5.0)
c3.upper_bound = 150.0

# Solve the problem
lp.simplex()

# Print results
if lp.status == GLPK.LP_OPT:
    print(f"Status: Optimal")
    print(f"Objective value: {lp.obj.value:.2f}")
    print(f"x1 = {x1.value:.2f}")
    print(f"x2 = {x2.value:.2f}")
elif lp.status == GLPK.LP_FEAS:
    print(f"Status: Feasible, but not necessarily optimal")
    print(f"Objective value: {lp.obj.value:.2f}")
    print(f"x1 = {x1.value:.2f}")
    print(f"x2 = {x2.value:.2f}")
else:
    print(f"Status: Solver terminated with status code: {lp.status}")

# Clean up
lp.delete()

view raw JSON →