Could your organization benefit from working with a Gurobi Alliance partner? That depends on your team’s in-house proficiency with mathematical optimization—as well as their bandwidth. To identify your team’s proficiency level, you can start by asking the following questions:
1. Does my team know how to build mathematical optimization models?
More and more, our clients have the in-house capacity to develop mathematical optimization models. They may not have the knowledge to develop the most elegant models, but they can get the process started.
2. Does my team employ optimization best practices?
While many firms have team members with the capacity to develop mathematical optimization models, fewer have personnel with the depth of knowledge and level of experience necessary to apply evolving industry best practices to their work.
In addition, some firms have team members with limited exposure and experience with mathematical optimization—where they can do more harm than good (i.e., the resulting model becomes overly complicated or does not use best practices). An outside expert who has devoted their career to best practices in mathematical optimization can help you know which approaches to apply and when, and how to appropriately test models for correctness.
3.Does my team study solver behavior?
A few firms have distinctly talented in-house teams who can develop a mathematical optimization model, understand the best techniques for validating that the model is correct, and appropriately deploy best practices. However, they may still need support in understanding how Gurobi (or another solver) interacts with the model.
Models with millions of variables and different ranges of data inputs can interface with a solver in unexpected ways. In this case, consider bringing in support from someone who is deeply familiar with the solver itself to help you understand what is happening and how to respond.
4. Does my team enhance solver-model interactions?
There is an art and science to writing a good mathematical optimization model—and a poorly written model may be the reason for long solve times. In addition to helping firms understand how their model and the solver are interacting, an outside partner can help their client improve the relationship between the two to make the results even more robust.
The Continuum of Support
Even teams with excellent optimization proficiency can use an extra pair of hands sometimes. Gurobi and our Gurobi Alliance partners are here for you when you need advice, or whenever you encounter a problem. No matter where you are on your optimization journey, our experts can come alongside and support you, with services that include:
Software-Level Technical Support
At the end of the day, Gurobi is a software company. Many of our customers have the mathematical programming skills necessary to write their own models. In this case, the Gurobi team provides top-tier user support in helping them address any software issues as they arise.
Model Fine Tuning
As we troubleshoot any software issues our customers encounter, Gurobi frequently offers insights to clients on how they can fine-tune their models. We provide this feedback as a courtesy. However, if we identify major gaps in the model, we can make recommendations on which of our partners to engage with and how.
If you need it, we can connect you with a Gurobi Alliance partner who can help you build a model from scratch. These partners can listen to your goals and objectives, map out your constraints, help you work through the tradeoffs and priorities, and develop a model that can guide your decision-making.
Model Review & Validation
Let’s say you have designed your model and incorporated it into your decisions and business practices. But suddenly, the model starts behaving oddly or producing unusual results. A consultant can review the model and identify the source of the issue.
For example, one of our customer’s models suddenly started generating unusual results. After a consultant worked with the client to review the model, they realized that the issue was in the data management and entry into the model (i.e., the firm’s “upstream” data practices). Identifying where the issue was located empowered the firm to course correct and stabilize their model and results.
Occasionally, a customer needs to completely rethink their approach to modeling. For example, one client’s model started running much more slowly than it had in the past as their business had grown. They then engaged Princeton Consultants to provide a fresh perspective on what could be going wrong.
Princeton helped them totally change their modeling approach, leveraging key Gurobi features that reduced solution times from 48 hours to 15 minutes on average. This allowed the client to experiment with different scenarios, opening up new business opportunities.
In addition to Gurobi’s expert support and our network of trusted partners, Gurobi’s research and development team is constantly searching for ways to improve our solver and provide our customers with additional functionality. For example, in response to one client’s needs, we expanded our recommendations regarding highly intensive monitoring systems for model corrections.
You Don’t Have to Go It Alone
To use an automotive metaphor, Gurobi makes an engine, but you are the driver. Depending on your skills and needs, our consulting partners can build the car for you or give you advice when the car you built yourself is not working the way you intended. Regardless of whether you are buying something off the shelf or building it yourself, working with Gurobi and our consulting partners can help you rethink what is possible with your mathematical modeling. You do not have to figure out all the answers on your own. We’re here to help.