Visit your regional NVIDIA website for local content, pricing, and where to buy partners specific to your country.
Decision Optimization
Achieve world-record accuracy and performance to handle large-scale problems with millions of constraints and variables, save time, and reduce costs.
Watch Video | Solution Brief | For Developers
NVIDIA® cuOpt™ is a groundbreaking optimization AI microservice excelling in fast decision optimization, including linear programming and mixed integer linear programming heuristics and VRP (Vehicle Routing Problems). It handles large-scale problems with millions of variables and constraints, enabling near real-time optimization, potentially saving millions of dollars. With 23 world-record benchmarks, cuOpt owns all of the recent world records on the largest routing benchmarks. It also accelerates state-of-the-art CPU linear programming techniques on Mittelmann’s benchmark, solving large-scale optimization problems with millions of variables and constraints.
cuOpt—a groundbreaking optimization AI microservice excelling in decision optimization—is now available.
cuOpt uses GPU-accelerated logistics solvers relying on heuristics, metaheuristics, and optimizations to calculate complex vehicle routing problems with a wide range of constraints. cuOpt can be deployed in any data center or cloud. With support for distance and time matrices with asymmetric patterns, it can be seamlessly integrated with popular map engines.
Faster on 60% of the Mittelmann LPOpt instances and more than 10X faster on 20% when compared to commercial SOTA.
Consistent speedups of 8X ~ 5,607X on multi-commodity flow problems.
Experience world-record performance, achieved across Li & Lim and Gehring & Homberger accuracy benchmarks.
Scale out to 15,000 routing tasks to facilitate computationally heavy use cases.
Route 1,000 packages in 10 seconds instead of 20 minutes (120X faster) with the same level of accuracy.
Rerun models and adjust for changes like inoperable vehicles, traffic and weather disruptions, and the addition of new orders—all within service-level agreement (SLA) time constraints.
Access a secure, production-ready microservice, as part of NVIDIA AI Enterprise, designed to deploy anywhere and accelerate time to value.
Use Cases
See how NVIDIA cuOpt supports industry use cases and jump-start your AI development with curated examples.
Resource allocation in complex supply chains involves efficiently distributing limited resources across tasks to maximize productivity and minimize costs. The challenge lies in numerous variables and real-time changes, necessitating rapid, optimal solutions to maintain operational agility. cuOpt AI agent allows you to talk to your supply chain data via LLM NIM and optimize your resource allocation.
Challenges in job scheduling optimization lie in maximizing productivity and minimizing idle time across complex systems with varying constraints. This process requires adapting to dynamic workloads and unexpected changes while maintaining overall system efficiency. With GPU acceleration, cuOpt enables businesses to make data-driven job scheduling decisions, optimizing for factors like deadlines, priorities, and resource availability.
Stock allocation in finance involves distributing investment capital across securities to optimize portfolio performance. The challenge lies in balancing risk and return objectives while considering market volatility, economic indicators, and investor preferences. Complexity arises from numerous potential combinations and the need for swift adaptation to changing market conditions.
Fleets of trucks are dispatched from distribution centers to deliver orders to stores and end customer addresses. cuOpt reduces miles driven and accelerates delivery times with the same level of accuracy.
A service provider is dispatched to fulfill a list of different service requests. These are allocated a certain amount of time per service request and the time can vary according to request. For example, a telecommunications technician is assigned to a customer’s home to install a router and then to a different home to install a data cable. cuOpt enables technicians to be equipped with all required tools before departure and complete the tasks based on optimized routes.
For the inbound and outbound transportation of goods and vehicles with long-haul fleets, it’s critical to optimize scheduling and route planning. cuOpt allows users to factor in the number of available pilots, drivers, or ships to recommend the most optimized route and schedule.
Use the right tools and technologies to take logistics optimization projects from development to production.
Experience cuOpt through a UI-based portal for exploring and prototyping with NVIDIA-managed endpoints, available for free through NVIDIA's API catalog.
Access NVIDIA-hosted infrastructure and guided hands-on labs that include step-by-step instructions and examples, available for free on NVIDIA LaunchPad.
Get a free license to try NVIDIA AI Enterprise in production for 90 days using your existing infrastructure.
Kawasaki Heavy Industries, Ltd. is a manufacturing company that’s been building large machinery for more than a hundred years. With NVIDIA cuOpt and Jetson Orin™, Kawasaki transformed its track maintenance and inspection capabilities.
In this self-paced course from the NVIDIA Deep Learning Institute, you’ll work through a common vehicle routing optimization problem and learn how to preprocess input data for use by cuOpt, composing variants of the problem that reflect real-world business constraints.
Have an accelerated optimization project? Get access to the route optimization workflow with a free curated lab in NVIDIA LaunchPad, which includes a step-by-step guide and ready-to-use software, sample data, and applications.
In this hands-on lab, learn how to use the NVIDIA cuOpt cloud service to find the most optimal routes for a heterogeneous fleet of vehicles making deliveries, pickups, dispatching jobs, and more.
Find out how organizations are driving greater efficiency, saving money, and boosting revenue and customer satisfaction with real-time route optimization.
Learn how, with NVIDIA Metropolis, Omniverse™, cuOpt, and Isaac™ for robot perception, it’s possible to create an end-to-end strategy for fully automating logistically complex co-bot spaces.
Watch how organizations can overcome operations complexities and deliver AI factories at extraordinary scale with an AI planner built with LLM NIMs, NVIDIA NeMo Retriever NIMs, and a cuOpt NIM.
Next Steps
Explore everything you need to start developing with NVIDIA cuOpt, including the latest documentation, tutorials, technical blogs, and more.
Talk to an NVIDIA product specialist about moving from pilot to production with the security, API stability, and support of NVIDIA AI Enterprise.