Artificial intelligence manages to predict the fate of planetary systems and to say which ones are most likely to survive. It is possible thanks to the model called Spock (Stability of Planetary Orbital Configurations Klassifier) published in the journal of the American Academy of Sciences (Pnas) by research led by Daniel Tamayo, of the American Princeton University. "Being able to identify which planetary systems are stable and which unstable is a fascinating and difficult problem, "Tamayo said.
To ensure that a planetary system is stable, astronomers must calculate the movements of the planets that populate it and interact with each other for billions of years, checking the stability of every possible configuration, a computationally prohibitive undertaking. However, it is an important calculation because it helps to identify which planetary systems are destined to become stable like ours and to survive.
To overcome the problem, Tamayo has decided to speed up the process by combining models of planetary interactions with machine learning methods. This allows the rapid elimination of huge unstable orbital configurations and calculations that would have required tens of thousands of hours in this way can be performed in minutes.
Instead of simulating a given configuration for a billion orbits, the Tamayo model simulates all possible configurations for 10,000 orbits, which only takes a fraction of a second. From this the 10 main dynamics of the system are calculated. Finally, a machine learning algorithm is used to predict from these 10 dynamics if, by letting it go for a billion orbits, the system remains stable, living long and prospering.
Artificial intelligence predicts the fate of planetary systems
2020-07-17T01:06:40.324Z
Says which ones are most likely to survive (ANSA)Artificial intelligence manages to predict the fate of planetary systems and to say which ones are most likely to survive. It is possible thanks to the model called Spock (Stability of Planetary Orbital Configurations Klassifier) published in the journal of the American Academy of Sciences (Pnas) by research led by Daniel Tamayo, of the American Princeton University. "Being able to identify whic...