Python API Reference

Modern Python bindings for the Routing library with automatic attribute enabling.

---

Quick Start

import routing
from routing.constants import Attribute

routing.init()

# Create problem
problem = routing.Problem()

# Add depot
depot = problem.add_depot(0)
depot.add_attribute(Attribute.GEONODE, x=0, y=0)

# Add client
client = problem.add_client(1)
client.add_attribute(Attribute.GEONODE, x=10, y=20)
client.add_attribute(Attribute.CONSUMER, demand=15)

# Add vehicle
vehicle = problem.add_vehicle(0)
vehicle.add_attribute(Attribute.STOCK, capacity=100)

# Solve (attributes auto-enabled!)
solution = routing.solve(problem, "ga", timeout=30)
print(f"Cost: {solution.cost}")

---

Attributes

Add attributes using the Attribute enum from routing.constants:

from routing.constants import Attribute

client.add_attribute(Attribute.GEONODE, x=10, y=20)
client.add_attribute(Attribute.CONSUMER, demand=5)
client.add_attribute(Attribute.RENDEZVOUS, open=0, close=100)
client.add_attribute(Attribute.SERVICE_QUERY, service_time=10)

Enum	Parameters	Description
Attribute.GEONODE	x, y	Coordinates
Attribute.CONSUMER	demand	Client demand
Attribute.STOCK	capacity	Vehicle capacity
Attribute.RENDEZVOUS	open, close	Time window
Attribute.SERVICE_QUERY	service_time	Service duration
Attribute.PROFITER	profit	Profit value (TOP)
Attribute.PICKUP	demand	Pickup demand
Attribute.DELIVERY	demand	Delivery demand

Attributes are automatically enabled when added—no manual setup required.

---

Problem Class

Creating a Problem

problem = routing.Problem()

Adding Entities

depot = problem.add_depot(0)
client = problem.add_client(1)
vehicle = problem.add_vehicle(0)

Properties

Property	Description
num_clients	Number of clients
num_vehicles	Number of vehicles
num_depots	Number of depots

---

Entity Methods

All entities (Client, Depot, Vehicle) share these methods:

from routing.constants import Attribute

# Add attribute (named args recommended)
entity.add_attribute(Attribute.GEONODE, x=10, y=20)

# Check if attribute exists
if entity.has_attribute(Attribute.GEONODE):
    print(f"Location: ({entity.x}, {entity.y})")

# Get entity ID
id = entity.get_id()

Entity Properties

Client:

Property	Attribute	Description
x, y	GeoNode	Coordinates
demand	Consumer	Demand
tw_open, tw_close	Rendezvous	Time window
service_time	ServiceQuery	Service duration
profit	Profiter	Profit value

Vehicle:

Property	Attribute	Description
capacity	Stock	Capacity

---

Solving

Simple Solve

solution = routing.solve(problem, "ga", timeout=30)

if solution:
    print(f"Cost: {solution.cost}")
    print(f"Feasible: {solution.is_feasible}")

With Progress Callback

def on_improvement(solution, cost):
    print(f"New best: {cost:.2f}")

solution = routing.solve_with_callback(
    problem, on_improvement, "ga", timeout=30
)

Manual Solver Control

solver = routing.create_solver("ga", problem)
solver.set_param_int("iterMax", 10000)
solver.solve(60.0)

solution = solver.get_solution()
print(f"Optimal: {solver.is_optimal()}")

---

Solution Class

Property	Description
cost	Total cost
num_tours	Number of routes
is_feasible	All clients served
unserved	Unserved client IDs

Getting Routes

for tour in solution.get_tours():
    print(f"Route: {tour.get_client_ids()}")
    print(f"Cost: {tour.cost}")

---

Loading Files

# Solomon format (CVRPTW)
problem = routing.load_solomon("c101.txt")

# TSPLIB format (CVRP)  
problem = routing.load_tsplib("A-n32-k5.vrp")

---

Discovery Functions

routing.list_solvers()      # Available solvers
routing.list_attributes()   # Available attributes
routing.get_problem_types() # Problem type requirements

---

Examples

CVRP

import routing
from routing.constants import Attribute

routing.init()

problem = routing.Problem()

# Depot
depot = problem.add_depot(0)
depot.add_attribute(Attribute.GEONODE, x=0, y=0)

# Clients
for i in range(1, 11):
    client = problem.add_client(i)
    client.add_attribute(Attribute.GEONODE, x=i*10, y=i*10)
    client.add_attribute(Attribute.CONSUMER, demand=5+i)

# Vehicles
for v in range(3):
    vehicle = problem.add_vehicle(v)
    vehicle.add_attribute(Attribute.STOCK, capacity=50)

# Solve
solution = routing.solve(problem, "ga", timeout=30)
print(f"Cost: {solution.cost:.2f}")

CVRPTW

from routing.constants import Attribute

problem = routing.Problem()

# Depot with time window
depot = problem.add_depot(0)
depot.add_attribute(Attribute.GEONODE, x=0, y=0)
depot.add_attribute(Attribute.RENDEZVOUS, open=0, close=1000)

# Client with time window
client = problem.add_client(1)
client.add_attribute(Attribute.GEONODE, x=20, y=20)
client.add_attribute(Attribute.CONSUMER, demand=10)
client.add_attribute(Attribute.RENDEZVOUS, open=0, close=100)
client.add_attribute(Attribute.SERVICE_QUERY, service_time=10)

# Vehicle
vehicle = problem.add_vehicle(0)
vehicle.add_attribute(Attribute.STOCK, capacity=50)

solution = routing.solve(problem, "alns", timeout=60)

---

Visualization

The library includes built-in visualization support using matplotlib.

Note

Requires matplotlib: pip install matplotlib

Plot Solution

from routing.visualization import plot_solution

# Basic plot
fig, ax = plot_solution(problem, solution)

# Customize
plot_solution(problem, solution,
              show_demand=True,
              show_arrows=True,
              title="My CVRP Solution",
              save_path="solution.png")

Plot Problem

from routing.visualization import plot_problem

# Show client locations and demands
fig, ax = plot_problem(problem, show_demand=True)

Plot Convergence

from routing.visualization import plot_convergence

# Collect costs during solving
costs = []
def callback(solution, cost):
    costs.append(cost)

solution = routing.solve_with_callback(problem, callback, "ga", 30)

# Plot convergence curve
plot_convergence(costs, title="GA Convergence")

Compare Solvers

from routing.visualization import compare_solutions

solutions = {
    "GA": routing.solve(problem, "ga", timeout=30),
    "ALNS": routing.solve(problem, "alns", timeout=30),
    "VNS": routing.solve(problem, "vns", timeout=30),
}

# Side-by-side comparison
compare_solutions(problem, solutions)

Gantt Chart (Time Windows)

For VRPTW problems, visualize the schedule with time windows:

from routing.visualization import plot_gantt

# Plot schedule showing time windows and service times
fig, ax = plot_gantt(problem, solution)

# Customize
plot_gantt(problem, solution,
           show_travel=True,     # Show travel time bars
           show_tw_bounds=True,  # Show time window boundaries
           speed=1.0,            # Travel speed for time calculation
           title="Vehicle Schedules")

The Gantt chart shows:

• Gray bars: Time window constraints [open, close]

• Colored bars: Service time at each client

• Thin bars: Travel time between clients (optional)

Visualization Options

Parameter	Default	Description
show_labels	True	Show client ID labels
show_demand	False	Show demand values
show_arrows	True	Show direction arrows
show_legend	True	Show route legend
route_width	1.5	Line width for routes
client_size	100	Marker size for clients
depot_size	200	Marker size for depot
figsize	(10, 8)	Figure size
save_path	None	Path to save figure

---

See Also

• Attributes - Detailed attribute documentation

• Solvers - Solver comparison and tuning

• Visualization Notebook - Interactive examples