Stay organized with collections
Save and categorize content based on your preferences.
One of the most common optimization tasks is vehicle routing, in which the
goal is to find the best routes for a fleet of vehicles visiting a set of
locations. Usually, "best" means routes with the least total distance or cost.
Here are a few examples of routing problems:
A package delivery company wants to assign routes for drivers to make deliveries.
A cable TV company wants to assign routes for technicians to make residential
service calls.
A ride-sharing company wants to assign routes for drivers to pick up and drop
off passengers.
The most famous routing problem is the Traveling Salesperson Problem (TSP):
find the shortest route for a salesperson who needs to visit customers at
different locations and return to the starting point. A TSP can be represented
by a graph, in which the nodes correspond to the locations, and the edges (or
arcs) denote direct travel between locations. For example, the graph below shows
a TSP with just four locations, labeled A, B, C, and D.
The distance between any two locations is given by the number next to the edge
joining them.
By calculating the distances of all possible routes, you can see that the
shortest route is ACDBA, for which the total distance is 35 + 30 + 15 + 10 = 90.
The problem gets harder when there are more locations. In the example above,
there are just six routes. But if there are ten locations (not counting the
starting point), the number of routes is 362880.
For 20 locations, the number jumps to 2432902008176640000.
An exhaustive search of all possible routes would be guaranteed to find the
shortest, but this is computationally intractable for all but small sets of
locations. For larger problems, optimization techniques are needed to
intelligently search the solution space and find an optimal (or near-optimal)
solution.
A more general version of the TSP is the vehicle routing problem (VRP),
in which there are multiple vehicles. In most cases, VRPs have constraints: for
example, vehicles might have capacities for the maximum weight or volume of items they
can carry, or drivers might be required to visit locations during specified time windows
requested by customers.
OR-Tools can solve many types of VRPs, including the following:
VRP with time windows, where the vehicles must visit the locations in
specified time intervals.
VRP with resource constraints, such as space or personnel
to load and unload vehicles at the depot (the starting point for the routes).
VRP with dropped visits, where the vehicles aren't required to
visit all locations, but must pay a penalty for each visit that is dropped.
Limitations on solving vehicle routing problems
Vehicle routing problems are inherently intractable: the length of time it takes
to solve them grows exponentially with the size of the problem. For sufficiently
large problems, it could take OR-Tools (or any other routing software) years to
find the optimal solution. As a result, OR-Tools sometimes returns solutions
that are good, but not optimal. To find a better solution, change the search
options for the solver. See Changing the search strategy
for an example.
[[["Easy to understand","easyToUnderstand","thumb-up"],["Solved my problem","solvedMyProblem","thumb-up"],["Other","otherUp","thumb-up"]],[["Missing the information I need","missingTheInformationINeed","thumb-down"],["Too complicated / too many steps","tooComplicatedTooManySteps","thumb-down"],["Out of date","outOfDate","thumb-down"],["Samples / code issue","samplesCodeIssue","thumb-down"],["Other","otherDown","thumb-down"]],["Last updated 2025-01-22 UTC."],[[["\u003cp\u003eVehicle routing focuses on finding optimal routes for vehicles to visit locations, minimizing distance or cost, with applications like delivery services and ride-sharing.\u003c/p\u003e\n"],["\u003cp\u003eThe Traveling Salesperson Problem (TSP) seeks the shortest route for one vehicle to visit all locations and return to the starting point, becoming computationally complex with more locations.\u003c/p\u003e\n"],["\u003cp\u003eOR-Tools addresses various Vehicle Routing Problems (VRPs), including TSP, VRP with capacity and time constraints, and VRP with resource limitations and potential dropped visits.\u003c/p\u003e\n"],["\u003cp\u003eSolving VRPs can be computationally challenging, with solution times increasing exponentially with problem size, sometimes resulting in good but not optimal solutions from OR-Tools.\u003c/p\u003e\n"],["\u003cp\u003eWhile specialized solvers like Concorde excel at large TSPs, OR-Tools provides a more versatile platform for routing problems with constraints beyond the basic TSP.\u003c/p\u003e\n"]]],["Vehicle routing problems (VRPs) involve finding optimal routes for vehicles visiting locations, often minimizing total distance or cost. The Traveling Salesperson Problem (TSP) is a specific VRP with one vehicle. Solving these problems involves searching possible routes, with complexity increasing exponentially with more locations. OR-Tools, a routing solver, addresses various VRPs, including those with capacity, time, resource, and penalty constraints. While not always finding the absolute optimal solution, OR-Tools provides good solutions and users can adjust search strategies for better results.\n"],null,["| **Note:** While the routing solver in Google's OR-Tools is free, users who need an industrial-class service can use the [Google Maps Platform Route Optimization API](https://developers.google.com/maps/documentation/route-optimization). Check out the [source code for a sample web application](https://github.com/googlemaps/js-route-optimization-app).\n\nOne of the most common optimization tasks is *vehicle routing*, in which the\ngoal is to find the best routes for a fleet of vehicles visiting a set of\nlocations. Usually, \"best\" means routes with the least total distance or cost.\nHere are a few examples of routing problems:\n\n- A package delivery company wants to assign routes for drivers to make deliveries.\n- A cable TV company wants to assign routes for technicians to make residential service calls.\n- A ride-sharing company wants to assign routes for drivers to pick up and drop off passengers.\n\nThe most famous routing problem is the *Traveling Salesperson Problem* (TSP):\nfind the shortest route for a salesperson who needs to visit customers at\ndifferent locations and return to the starting point. A TSP can be represented\nby a graph, in which the nodes correspond to the locations, and the edges (or\narcs) denote direct travel between locations. For example, the graph below shows\na TSP with just four locations, labeled A, B, C, and D.\nThe distance between any two locations is given by the number next to the edge\njoining them.\n\nBy calculating the distances of all possible routes, you can see that the\nshortest route is ACDBA, for which the total distance is `35 + 30 + 15 + 10 = 90`.\n\nThe problem gets harder when there are more locations. In the example above,\nthere are just six routes. But if there are ten locations (not counting the\nstarting point), the number of routes is 362880.\nFor 20 locations, the number jumps to 2432902008176640000.\nAn exhaustive search of all possible routes would be guaranteed to find the\nshortest, but this is computationally intractable for all but small sets of\nlocations. For larger problems, optimization techniques are needed to\nintelligently search the solution space and find an optimal (or near-optimal)\nsolution.\n\nA more general version of the TSP is the vehicle routing problem (VRP),\nin which there are multiple vehicles. In most cases, VRPs have constraints: for\nexample, vehicles might have capacities for the maximum weight or volume of items they\ncan carry, or drivers might be required to visit locations during specified time windows\nrequested by customers.\n\nOR-Tools can solve many types of VRPs, including the following:\n\n- [Traveling Salesperson Problem](/optimization/routing/tsp), the classic routing problem in which there is just one vehicle.\n- [Vehicle routing problem](/optimization/routing/vrp), a generalisation of the TSP with multiple vehicles.\n- [VRP with capacity constraints](/optimization/routing/cvrp), in which vehicles have maximum capacities for the items they can carry.\n- [VRP with time windows](/optimization/routing/vrptw), where the vehicles must visit the locations in specified time intervals.\n- [VRP with resource constraints](/optimization/routing/cvrptw_resources), such as space or personnel to load and unload vehicles at the depot (the starting point for the routes).\n- [VRP with dropped visits](/optimization/routing/penalties), where the vehicles aren't required to visit all locations, but must pay a penalty for each visit that is dropped.\n\nLimitations on solving vehicle routing problems\n\nVehicle routing problems are inherently intractable: the length of time it takes\nto solve them grows exponentially with the size of the problem. For sufficiently\nlarge problems, it could take OR-Tools (or any other routing software) years to\nfind the optimal solution. As a result, OR-Tools sometimes returns solutions\nthat are good, but not optimal. To find a better solution, change the search\noptions for the solver. See [Changing the search strategy](/optimization/routing/tsp#search_strategy)\nfor an example.\n| **Note:** We should add that there are other solvers, such as [Concorde](http://www.math.uwaterloo.ca/tsp/concorde.html), dedicated to solving very large TSPs to optimality, which surpass OR-Tools in that area. However, OR-Tools provides a better platform for solving more general routing problems that contain constraints beyond those of a pure TSP."]]
