At our latest webinar with Gurobi, The Next Generation of Cloud Optimization Development, we took a deep dive into how users can maximize the power of Gurobi using our cloud based platform. Here are some of the most common questions we received:
To try the Gurobi optimizer in Optilogic, do I need an account? How much does it cost?
Yes, you will need an account, but it's free to start! We have a free community tier that gives you access to a little bit of everything so you can try out our entire platform.
Are there any limitations on model size when using Optilogic and Gurobi?
The great thing about the cloud is the scalability of the resources at hand. You have the choice with Optilogic to run a model on a 4 core, 8GB machine, all the way up to a 16 core, 128GB machine, and many configurations in between. So we can easily scale up to meet the needs of very complex and large problems.
Is it possible to run a job remotely if I have additional 3rd party dependencies (e.g. data connector libraries)?
We offer a couple ways to modify your environment to your needs (example: add dependencies or other libraries), and we manage the creation and management of containers for running your model in our backend execution system. You have access to all of this through our APIs. Our site documentation has more in depth information at docs.optilogic.com.
Is Optilogic targeted towards the data scientist and MIP community or software developers? How accessible would it be for an experienced MIP modeler to learn to use Optilogic for modeling?
Optilogic has two main audiences: those who build the model formulations and those who develop solutions around the model (app builders). For modelers, they can dive into model formulations and use our platform as a companion to their current toolset. We hope Optilogic feels like that perfect tool for your skillset. If there are models that you need help building, we can also help coach you through this process. If you are a developer teamed up with a DataSci or MIP, you will find our APIs clean and easy to use for running the models in your application.
How can users deploy applications?
Optilogic users can deploy applications internally and externally. These applications can utilize our clean set of APIs to connect to models and data that live in our system, and in doing so, access our back-end scalability by default. You can do these yourself, or if you need a hand, give us a buzz and we’ll make sure you are successful. (we’re a startup! We just want you to have success using the tool.)
What’s the difference between "Run it now" and "Run as a job"?
‘Run it now’, runs the model in your current operating environment in the cloud. This is sharing resources with everything you are currently doing on a preset server for you. ‘Run as job’ allows you to spin up a new environment server that is targeted just for running heavy processing. You can select the size you need to run and fire off the job. There’s a very minor time overhead to this, so sometimes running in your current environment is quicker and easier, plus gives you the full debug and variable tracking capabilities. While running in a job assumes you need nothing but pure horsepower.
What are the tradeoffs between using Andromeda for cloud solver scalability and using Gurobi Cloud platform for cloud solver scalability?
The main difference is ease of use in elastic scalability and licensing headaches. With Optilogic you can spin up hundreds of nodes without needing to change any configuration, you simply need to run the models. There are no firewalls, no resource controls or licensing to figure out. We have removed the need to manage your account or a bank of servers in a cluster because we do it for you.
How does Optilogic compare to Gurobi remote services?
Gurobi remote services is a great resource that allows you to manage a cluster of machines. The main difference here is that there is nothing that you need to manage and you get access to elastic scalability. To take advantage of running hundreds of solves at once, or changing the machine resources that you are running on, you simply need to run the model and select your resource configuration. We take care of the rest.
If you want to learn more about our partnership with Gurobi, you can download the webinar here.