Flipy

Flipy is a Python linear programming interface library, originall developed by FreeWheel. It currently supports Gurobi and CBC as the backend solver.

To use Gurobi, make sure you have a Gurobi license file, and gurobipy is installed in your Python environment. You can find details from Gurobi’s documentation.

Flipy requires Python 3.6 or newer.

Installation

Flipy can be installed with setuptools.

$ git clone https://github.freewheel.tv/linear/flipy
$ cd flipy
$ python setup.py install

Quickstart

Here is a simple example for Flipy:

import flipy

# 1 <= x <= 3.5
x = flipy.LpVariable('x', low_bound=1, up_bound=3.5)
# 2 <= y <= 4
y = flipy.LpVariable('y', low_bound=2, up_bound=4)

# 5x + y <= 12
lhs = flipy.LpExpression('lhs', {x: 2.5, y: 1})
rhs = flipy.LpExpression('rhs', constant=12) 
constraint = flipy.LpConstraint(lhs, 'leq', rhs)

# maximize: 3x + 2y
objective = flipy.LpObjective('test_obj', {x: 3, y: 2}, sense=flipy.Maximize)
problem = flipy.LpProblem('test', objective, [constraint])

solver = flipy.CBCSolver()
status = solver.solve(problem)

Get the solution

After solving, a status is returned to indicate whether the solver has found a optimal solution for the problem:

print(status)
# <SolutionStatus.Optimal: 1>

The objective value can be retrieved with objective.evaluate():

print(objective.evaluate())
# 17.6

The value of variables can be retrieved with .evaluate() as well:

print(x.evaluate())
# 3.2
print(y.evaluate())
# 4.0