The International Space Station (ISS) depends on solar energy, but keeping the lights on in space is easier said than done.

During every second of its 92-minute orbit of Earth, the station’s solar panels must be precisely positioned to maximize energy capture without casting uneven shadows on the longerons — the long, thin, tethering beams that hold the panels to the station. Longerons are highly sensitive to temperature, and uneven shadows can cause them to warp and snap.

NASA relies on a complex algorithm to turn motors that reposition the solar collectors and ensure safe, efficient energy capture aboard the ISS, but they wanted to determine whether the algorithm could be optimized to deliver further gains. Together with Topcoder, NASA launched the ISS Longeron Challenge to put their algorithm to the test by developing multiple new approaches.

First, the Topcoder Community developed a high performance Java-based model of the station to assist competitors in their pursuit of a new algorithmic solution. By using this model, competitors were able to test their solutions at home in under 15 seconds. Competitors then submitted their code and solutions to Topcoder for automated, online testing. Results were computed in real time during the challenge and posted to a public leaderboard to drive competition.

More than 4,000 citizen scientists registered for the three-week challenge, and 459 competitors submitted over 2,000 possible solutions. The winner of the challenge was a data scientist with no prior experience solving problems related to space or solar capture.

In the end, the winning solution delivered by Topcoder provided an average output similar to NASA's original solution — but with new motor configurations. While the top solution didn't deliver more efficient solar capture, the challenge was deemed a success because it enabled NASA to validate the original algorithm, as well as see alternate approaches to the problem in action.

Data Scientists
459
Submissions
2010
Prize Pool
$30,000
Duration
3 Weeks