JijZept Solver

import jijzept_solver
import jijmodeling as jm

# Distance data for 5 cities (large diagonal values keep self-loops out)
num_cities = 5
distances = [
    [float(abs(i - j)) for j in range(num_cities)]
    for i in range(num_cities)
]
for city in range(num_cities):
    distances[city][city] = 1_000_000.0

@jm.Problem.define("TSP", sense=jm.ProblemSense.MINIMIZE)
def tsp(problem: jm.DecoratedProblem):
    d = problem.Float(ndim=2)  # Distances between cities
    N = problem.NamedExpr(d.len_at(0))  # Number of cities
    x = problem.BinaryVar(shape=(N, N))  # Whether to travel from city i to city j

    # Objective function: minimize total travel distance
    problem += jm.sum(d[i, j] * x[i, j] for i in N for j in N)
    # No tour constraints here -- the structure declaration tells the solver

# Create OMMX instance
instance = tsp.eval({"d": distances})

# Declare "x has the tour-route (TSP) structure" in one line and solve
solution = jijzept_solver.solve(
    instance,
    structure=jijzept_solver.Structure.tsp("x"),
    time_limit=10.0,
)
print(f"Total travel distance: {solution.objective}")