Logo
23.04

NVIDIA cuOpt

  • Overview
  • Setup
    • Docker
      • Prerequisites for running container
      • How to access the container
      • NVIDIA cuOpt Server
      • NVIDIA cuOpt through Juptyer Notebook
      • NVIDIA cuOpt Python
      • Helm Chart

Server

  • Server Overview
  • Server API Reference
    • Server
  • Example
    • Running cuOpt Server for CVRPTW

Python

  • Python Overview
  • Python API Reference
    • Routing
      • Data Model
        • Objective
        • DataModel
      • Solver Settings
        • InitialStrategy
        • SolutionStrategy
        • Scope
        • SolverSettings
      • Solve
        • Solve()
      • Assignment
        • Assignment
    • Distance engine
      • Waypoint Matrix
        • WaypointMatrix
  • Example
    • CVRPTW example over Homberger dataset

GETTING STARTED

  • Getting Started
    • cuOpt Basics
      • Overview
      • TSP, VRP, and PDP
      • What is NP-Hard?
      • The Necessity for Heuristics
      • How cuOpt Solves the Problem
      • GPUs Unleash Massive Parallel Computing Capabilities
      • Climbers
      • Approaches to Using cuOpt
      • Setup
      • cuOpt Memory Usage
    • cuOpt Supported Features
    • cuOpt Troubleshooting
      • Docker Sanity Check
      • Docker Not Installed or the Wrong Version
      • Docker Does Not Start
      • cuOpt Service Not Starting
      • cuOpt Service Not Responding
      • cuOpt Service Monitoring
      • How to Check if My Configuration Is Valid for a cuOpt Variant (Constraints, Inputs) When Using the cuOpt Service?
      • Common Misconfigurations Reported by the cuOpt Service, and How to Fix Them (Missing Data, Incompatible Constraints)
    • cuOpt FAQ
      • cuOpt resource estimates (how large a problem can I run with xyzconstraints).
      • How to set up multi-GPU cuOpt on a single server
      • How to set up high-availability service using Kubernetes + Helm Charts
      • How do I solve (some realistic problem) -> this is probably a decision tree that points to some variant of CVRP or PDP with constraints that match their needs but do not over-constraint the problem.
      • Do I need a GPU to use cuOpt?
      • What CUDA versions are supported?
      • Can I deploy cuOpt to my cloud environment?
      • Is CuOpt compatible with the new NVIDIA Hopper DPX instructions?
      • What is the memory consumption compared to the scale of data/model settings (for example, locations or number of climbers).
      • Is there a way to change the default port of cuOpt server?
      • Installation error - when installing NVIDIA driver on Ubuntu 20.04:
      • Is cuOpt supported on NVIDIA H100/Hopper architecture?
      • Is there any lower bound as Mixed integer gives a solution and lower bound?
      • How do we account for dynamic changing constraints while the solver is executing?
      • How are order cancellations handled in cuOpt?
      • Is there a C++ SDK to the cuOpt?
      • How many climbers do I need?
      • Can cuOpt be combined with Reinforcement Learning?
      • If the time limit is set to X seconds and the solver completes earlier, will it still run for X seconds?
      • Do we need to normalize the data when creating a time window matrix?
      • How does cuOpt check for collisions in NVIDIA Isaac Sim?
      • While setting certain customer nodes as break points, using add_break_dimensions(), there is an error where the result has additional locations.
      • How can we set the demand of a customer node higher than the capacity of the vehicle?
      • How to ensure that cuOpt does not assign high priority orders to dummy vehicles or technicians (that were added to get feasible solutions)?
      • When should I use the microservice vs Python API?
      • While calling API /cuopt/get_optimized_routes, we get an error {“detail”:”Failed to find solution with given constraints”}. Is it possible to know due to which constraint this failed?
      • How can I add multiple capacity information for a specific vehicle—height, widths and length, mileage, carbon footprint, and so on.
      • Is there a way to prevent vehicles from traveling along the same path in a waypoint graph, or is there a way to prevent more than one vehicle from visiting a location, or even that a location is only visited one time by a single vehicle?
      • Can I use my own data instead of Homberger’s instance data in cuOpt?
      • RuntimeError: cuOpt failure at file=/cuopt-build-utilities/cuopt/cpp/src/solution/get_solution.cu line=196: Truck capacity is smaller than volume.
      • The property num_of_climbers: int. What number is a typical value? What is the default value?
      • Travel time deviation—When using the same dataset, the travel timevaries by a couple of seconds in different runs, but the distance remains the same. How can travel time deviate in multiple runs on the same data and distance remains constant?
      • Null values in the cost matrix
      • I cannot model road conditions like traffic.
      • Can a break happen anywhere?
      • Time windows issues
      • What is the difference between set_min_vehicles and the horizontal loading feature?
      • Does cuOpt implement a topo sort for every route map?
      • How to choose various solution strategies?
      • Floating point vs integers when using the Set Task data server API call.
      • How do you compute the max driving that does not include wait time or service time?
      • When do I use set_vehicle_max_costs vs set_vehicle_max_times?
      • How do I calculate and adjust the optimal route while avoiding collisions using cuOpt and Issac Sim?
      • Is it possible to define constraints such as refrigerated vehicles required for certain orders?
      • Are there any plans to allow each vehicle to have its own maximum route distance so that mixed fleets with electric and traditional vehicles can be optimized?
      • What is the role of set_order_location()in pickup delivey combination?
      • How to model inputs for following scenarios?
      • How do we model the following scenario—Pick up from multiple different locations and deliver to a single customer.
      • We enabled the following “error_logging”: “verbose_mode”: true but still did not get any reason why it failed?
      • I know that the problem has a feasible solution, but cuOpt fails with infeasible. How to avoid this?
      • Where can I learn more about cuOpt?

RESOURCES

  • Resources
    • Examples and Cloud Scripts
    • Questions/Bugs/Feature-Request
    • Forum

RELEASE NOTES

  • Release Notes

Licenses

  • License
cuOpt
  • License
  • View page source

License¶

Please read https://docs.nvidia.com/cuopt/pdfs/NVIDIA-cuOpt-EULA.pdf for cuOpt

Previous

© Copyright 2021-2022, NVIDIA.

Built with Sphinx using a theme provided by Read the Docs.