What could be more complex than walking through an entirely new planet? NASA’s Jet Propulsion Laboratory (JPL) has sought the use of quantum computing in managing communications with its astronauts exploring the infinite universe. Through the use of the Azure Quantum system, JPL now communicates space missions through the Deep Space Network (DSN), a global network of large radio antennae located in California, Spain, and Australia that allows constant communication with spacecraft as the earth rotates.
With the intention of eventually incorporating a broader set of criteria, the Azure Quantum team built a solution for a variant of JPL’s scheduling problem with a constrained feature set. This is designed to speed up the whole process by reducing the need for lengthy negotiations.
Requests for space missions’ DSN antennae to be used in scheduling come with a lot of limits and a lot of computational power. When each spacecraft is visible to the antenna, all missions demand access for critical communication, resulting in hundreds of requests per week.
To create a schedule, the Microsoft team documented runtimes of two hours or more at the start of the project. The Microsoft team used Azure Quantum to cut the time needed to 16 minutes and a custom solution to roughly two minutes by using quantum-inspired optimization methods. Not only does JPL have the ability to construct a large number of candidate schedules in minutes rather than hours, but it also has the ability to be more flexible as the organization’s missions and expectations grow.
For most people, the scheduling problem at JPL and NASA’s deep space network may seem to appear as complex, but it is a typical problem in all businesses. This is a problem of job-shop scheduling, where jobs are assigned to resources as quickly as possible. Manufacturing supply chains, health care, transportation, and logistics are all examples of these issues.
Read more about the news here.
