JijModeling

2.4.1 verified Sat May 09 auth: no python

JijModeling is a Python library for mathematical modeling of optimization problems, with a focus on combinatorial optimization. It provides an intuitive API to define decision variables, constraints, and objective functions, and interfaces with solvers like JijZept. Current version 2.4.1, requires Python >=3.11, <4. Released regularly.

pip install jijmodeling

Common errors

error ModuleNotFoundError: No module named 'jijmodeling' ↓

cause jijmodeling is not installed or installed in wrong environment.

fix

Run pip install jijmodeling in the correct Python environment (Python >=3.11).

error AttributeError: module 'jijmodeling' has no attribute 'Model' ↓

cause In version 2.x, `Model` was renamed to `Problem`.

fix

Use jm.Problem instead of jm.Model.

error JijModelingError: Cannot use operator '>' with Expression objects ↓

cause Using Python comparison operators like `>` outside of Constraint definition.

fix

Wrap comparisons in jm.Constraint('name', expr) or use jm.Constraint(expr).

Warnings

breaking In version 2.0+, the API changed significantly. The old `jijmodeling` standalone model definition (e.g., `jm.Model`) is replaced by `jm.Problem`. Code from v1.x must be rewritten. ↓

fix Migrate from `jm.Model` to `jm.Problem`. Use `jm.BinaryVar` instead of `jm.Binary`. Check migration guide.

gotcha Variable indexing returns a placeholder, not the actual value. You cannot compare variables directly using Python comparison operators; use `==`, `<=`, `>=` only within constraints or objective expressions. ↓

fix Always use `jm.Constraint` for comparisons. Do not use `if x[0] == 1:` in code logic.

deprecated The `jm.pw` (piecewise linear) function is deprecated in 2.4.0+. Use `jm.Problem.add_piecewise` or custom linearization. ↓

fix Replace `jm.pw(...)` with `problem.add_piecewise(...)` or reformulate.

Imports

jm
wrong
```
from jijmodeling import *
```
correct
```
import jijmodeling as jm
```
Wildcard import pollutes namespace; use 'import jijmodeling as jm'.
Problem
wrong
```
from jijmodeling import Problem
```
correct
```
import jijmodeling as jm; problem = jm.Problem('name')
```
Direct import is fine, but typical usage via jm.Problem is recommended for clarity.

Quickstart

Define a simple binary optimization problem with variables, objective, and constraint.

import jijmodeling as jm

# Define variables
x = jm.BinaryVar('x', shape=(3,))
y = jm.BinaryVar('y', shape=(3,))

# Set problem
problem = jm.Problem('simple')
problem += x[0] + x[1] + x[2]
problem += jm.Constraint('c1', x[0] + y[0] == 1)

# Solve (requires JijZept token)
token = os.environ.get('JIJZEPT_TOKEN', '')
if token:
    import jijzept as jz
    sampler = jz.JijSASampler(token)
    result = sampler.sample(problem)
    print(result)
else:
    print('No JIJZEPT_TOKEN set. Problem defined:', problem)