A Machine Learns to Predict the Stability of Tightly Packed Planetary Systems

Abstract

The requirement that planetary systems be dynamically stable is often used to vet new discoveries or set limits on unconstrained masses or orbital elements. This is typically carried out via computationally expensive N-body simulations. We show that characterizing the complicated and multi-dimensional stability boundary of tightly packed systems is amenable to machine learning methods. We find that training an XGBoost machine learning algorithm on physically motivated features yields an accurate classifier of stability in packed systems. On the stability timescale investigated ($10^7$ orbits), it is 3 orders of magnitude faster than direct N-body simulations. Optimized machine learning classifiers for dynamical stability may thus prove useful across the discipline, e.g., to characterize the exoplanet sample discovered by the upcoming Transiting Exoplanet Survey Satellite (TESS). This proof of concept motivates investing computational resources to train algorithms capable of predicting stability over longer timescales and over broader regions of phase space.

arXiv commentAccepted in ApJ letters. 7 pages, 3 figures

References in corpus (19)

• XGBoost: A Scalable Tree Boosting System
• REBOUND: An open-source multi-purpose N-body code for collisional dynamics
• Dynamical Outcomes of Planet-Planet Scattering
• Architecture and Dynamics of Kepler's Candidate Multiple Transiting Planet Systems
• ANNz: estimating photometric redshifts using artificial neural networks
• WHFast: A fast and unbiased implementation of a symplectic Wisdom-Holman integrator for long term gravitational simulations
• Transit Timing Observations from Kepler: VII. Confirmation of 27 planets in 13 multiplanet systems via Transit Timing Variations and orbital stability
• Stability Limits in Extra-solar Planetary Systems
• The stability and fates of hierarchical two-planet systems
• Measurement of planet masses with transit timing variations due to synodic "chopping" effects
• Transit Timing Observations from Kepler: II. Confirmation of Two Multiplanet Systems via a Non-parametric Correlation Analysis
• Theory of Secular Chaos and Mercury's Orbit
• Chaotic Disintegration of the Inner Solar System
• Rapid dynamical chaos in an exoplanetary system
• Dynamical Stability of Imaged Planetary Systems in Formation: Application to HL Tau
• Consequences of an Eccentric Orbit for Fomalhaut b
• Dynamical behaviour of multiplanet systems close to their stability limit
• A simple scaling for the minimum instability time-scale of two widely spaced planets
• Spacing of Kepler Planets: Sculpting by Dynamical Instability

Cited by in corpus (39)

• Predicting the long-term stability of compact multiplanet systems
• Convergent Migration Renders TRAPPIST-1 Long-lived
• Warm terrestrial planet with half the mass of Venus transiting a nearby star
• Early 2017 observations of TRAPPIST-1 with $\text{Spitzer}$
• A Criterion for the Onset of Chaos in Systems of Two Eccentric Planets
• The path to instability in compact multi-planetary systems
• Separation of pulsar signals from noise with supervised machine learning algorithms
• Hill stability in the AMD framework
• New Constraints on Gliese 876 -- Exemplar of Mean-Motion Resonance
• A machine learns to predict the stability of circumbinary planets
• Exo-Milankovitch Cycles II: Climates of G-dwarf Planets in Dynamically Hot Systems
• Fly-by encounters between two planetary systems II: Exploring the interactions of diverse planetary system architectures
• Dynamical instabilities in systems of multiple short-period planets are likely driven by secular chaos: a case study of Kepler-102
• Fundamental limits from chaos on instability time predictions in compact planetary systems
• Using Deep Neural Networks to compute the mass of forming planets
• A Recommendation Algorithm to Predict Giant Exoplanet Host Stars Using Stellar Elemental Abundances
• Classifying Stars, Galaxies and AGN in CLAUDS+HSC-SSP Using Gradient Boosted Decision Trees
• Stability Constrained Characterization of Multiplanet Systems
• Predicting multiple planet stability and habitable zone companions in the TESS era
• Orbital stability of compact three-planet systems, I: Dependence of system lifetimes on initial orbital separations and longitudes
• Machine Learning Classification of Kuiper Belt Populations
• Follow the Water: Finding Water, Snow and Clouds on Terrestrial Exoplanets with Photometry and Machine Learning
• Machine learning prediction for mean motion resonance behaviour -- The planar case
• Architecture of three-planet systems predicted from the observed protoplanetary disk of HL Tau
• Orbital stability of compact three-planet systems, II: Post-instability impact behaviour
• A Pair of Planets Likely in Mean-motion Resonance From Gravitational Microlensing
• A hybrid approach for solving the gravitational N-body problem with Artificial Neural Networks
• Machine learning approach for mapping the stable orbits around planets
• Exploring the Dependence of Gas Cooling and Heating Functions on the Incident Radiation Field with Machine Learning
• A Bayesian neural network predicts the dissolution of compact planetary systems
• On the stability of planetary orbits in binary star systems I. The S-type orbits
• Secular Dynamics of Compact Three-Planet Systems
• The viewing angle in AGN SED models, a data-driven analysis
• ARDENT: A Python package for fast dynamical detection limits with radial velocities
• Orbital stability of compact three-planet systems III. The role of three-body resonances
• A Criterion for the Onset of Chaos in Compact, Eccentric Multiplanet Systems
• Advanced Techniques in Stability Analysis of Trans-Neptunian Objects
• Instability from high-order resonant chains in wide-separation massive planet systems
• The resilience of the sailboat stable region