GatorPy

GatorPy is an educationally motivated Linear Programming (LP) solver implemented in pure Python and NumPy. This project aims to demystify the internal workings of LP solvers by providing a simple, modular, and object-oriented framework. GatorPy transforms symbolic problem descriptions into slack form and solves them using the two-phase Simplex method.

See the below toy optimization example solved in GatorPy:

# Parameters
A = Parameter(np.array([[1,1],[1.2,0.5]]))
b = Parameter(np.array([1,1]))
c = Parameter(np.array([1.2,1]))

# Variables
y = Variable(2)

# Problem
problem = Problem({
    'maximize': c.T @ y,
    'subject to': [
        A @ y <= b,
        y <= 1,
        y >= 0
    ]
})

solution = problem.solve()
print(solution)
>>> (1.14, [0.71, 0.29], True)

The GatorPy solver supports a custom modeling language (as seen above) and is designed to be readable, extensible, and transparent for other students and practicioners embarking on similar projects

View the full GatorPy report at this link