Deep Reinforcement Learning for Efficient Scheduling of Ground-based Astronomical Observations

Ground-based astronomical observations face inherent challenges from weather changes, target visibility window constraints, and observational requirements. Enhancing the efficiency and effectiveness of telescope operations has long been a key objective for many observatories because of the high cost...

Full description

Saved in:
Bibliographic Details
Main Authors: Hai Cao, Shaoming Hu, Junju Du, Xu Chen, Shuqi Liu, Shuai Feng, Bo Zhang, Yuchen Jiang
Format: Article
Language:English
Published: IOP Publishing 2025-01-01
Series:The Astronomical Journal
Subjects:
Online Access:https://doi.org/10.3847/1538-3881/ade3dc
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849720779811323904
author Hai Cao
Shaoming Hu
Junju Du
Xu Chen
Shuqi Liu
Shuai Feng
Bo Zhang
Yuchen Jiang
author_facet Hai Cao
Shaoming Hu
Junju Du
Xu Chen
Shuqi Liu
Shuai Feng
Bo Zhang
Yuchen Jiang
author_sort Hai Cao
collection DOAJ
description Ground-based astronomical observations face inherent challenges from weather changes, target visibility window constraints, and observational requirements. Enhancing the efficiency and effectiveness of telescope operations has long been a key objective for many observatories because of the high cost of observational resources. In this study, we formalize observation scheduling as a time-dependent combinatorial optimization problem. To achieve this, we implement a pointer network with temporal attention that is capable of planning observations while accounting for time-varying factors such as moonlight interference, target altitude, and air mass, which impact the exposure time and image quality. To support the training of the deep neural network, we propose a scoring mechanism to evaluate the effectiveness of the observations, which is optimized through a refined REINFORCE algorithm with a baseline. Furthermore, an exposure time calculator and an equipment kinematic model are incorporated to dynamically estimate the time costs during the decision-making process. The simulation results demonstrated that the trained model significantly outperformed both manual scheduling and a greedy algorithm in terms of theoretical reward scores and the total number of scheduled targets. Observation experiments conducted using a dual-telescope system at Muztaga observatory further validated the superiority of our approach, demonstrating a 45.8% enhancement in the total signal-to-noise ratio across all observed targets and a 24.1% increase in the number of completed tasks under the same observing conditions.
format Article
id doaj-art-a6c80bc036284d639015e66db5890512
institution DOAJ
issn 1538-3881
language English
publishDate 2025-01-01
publisher IOP Publishing
record_format Article
series The Astronomical Journal
spelling doaj-art-a6c80bc036284d639015e66db58905122025-08-20T03:11:52ZengIOP PublishingThe Astronomical Journal1538-38812025-01-0117028810.3847/1538-3881/ade3dcDeep Reinforcement Learning for Efficient Scheduling of Ground-based Astronomical ObservationsHai Cao0https://orcid.org/0000-0001-9578-6613Shaoming Hu1https://orcid.org/0000-0003-3217-7794Junju Du2https://orcid.org/0000-0002-8571-8235Xu Chen3https://orcid.org/0000-0001-5603-7521Shuqi Liu4Shuai Feng5https://orcid.org/0000-0001-7288-7251Bo Zhang6Yuchen Jiang7Shandong Key Laboratory of Optical Astronomy and Solar-Terrestrial Environment, School of Space Science and Technology, Institute of Space Sciences, Shandong University , Weihai, Shandong 264209, People’s Republic of China ; husm@sdu.edu.cn; School of Mechanical, Electrical and Information Engineering, Shandong University , Weihai, Shandong 264209, People’s Republic of ChinaShandong Key Laboratory of Optical Astronomy and Solar-Terrestrial Environment, School of Space Science and Technology, Institute of Space Sciences, Shandong University , Weihai, Shandong 264209, People’s Republic of China ; husm@sdu.edu.cnShandong Key Laboratory of Optical Astronomy and Solar-Terrestrial Environment, School of Space Science and Technology, Institute of Space Sciences, Shandong University , Weihai, Shandong 264209, People’s Republic of China ; husm@sdu.edu.cnShandong Key Laboratory of Optical Astronomy and Solar-Terrestrial Environment, School of Space Science and Technology, Institute of Space Sciences, Shandong University , Weihai, Shandong 264209, People’s Republic of China ; husm@sdu.edu.cnShandong Key Laboratory of Optical Astronomy and Solar-Terrestrial Environment, School of Space Science and Technology, Institute of Space Sciences, Shandong University , Weihai, Shandong 264209, People’s Republic of China ; husm@sdu.edu.cnShandong Key Laboratory of Optical Astronomy and Solar-Terrestrial Environment, School of Space Science and Technology, Institute of Space Sciences, Shandong University , Weihai, Shandong 264209, People’s Republic of China ; husm@sdu.edu.cnShandong Key Laboratory of Optical Astronomy and Solar-Terrestrial Environment, School of Space Science and Technology, Institute of Space Sciences, Shandong University , Weihai, Shandong 264209, People’s Republic of China ; husm@sdu.edu.cnShandong Key Laboratory of Optical Astronomy and Solar-Terrestrial Environment, School of Space Science and Technology, Institute of Space Sciences, Shandong University , Weihai, Shandong 264209, People’s Republic of China ; husm@sdu.edu.cnGround-based astronomical observations face inherent challenges from weather changes, target visibility window constraints, and observational requirements. Enhancing the efficiency and effectiveness of telescope operations has long been a key objective for many observatories because of the high cost of observational resources. In this study, we formalize observation scheduling as a time-dependent combinatorial optimization problem. To achieve this, we implement a pointer network with temporal attention that is capable of planning observations while accounting for time-varying factors such as moonlight interference, target altitude, and air mass, which impact the exposure time and image quality. To support the training of the deep neural network, we propose a scoring mechanism to evaluate the effectiveness of the observations, which is optimized through a refined REINFORCE algorithm with a baseline. Furthermore, an exposure time calculator and an equipment kinematic model are incorporated to dynamically estimate the time costs during the decision-making process. The simulation results demonstrated that the trained model significantly outperformed both manual scheduling and a greedy algorithm in terms of theoretical reward scores and the total number of scheduled targets. Observation experiments conducted using a dual-telescope system at Muztaga observatory further validated the superiority of our approach, demonstrating a 45.8% enhancement in the total signal-to-noise ratio across all observed targets and a 24.1% increase in the number of completed tasks under the same observing conditions.https://doi.org/10.3847/1538-3881/ade3dcGround-based astronomyObservational astronomyAstronomical techniquesNeural networksOptical observation
spellingShingle Hai Cao
Shaoming Hu
Junju Du
Xu Chen
Shuqi Liu
Shuai Feng
Bo Zhang
Yuchen Jiang
Deep Reinforcement Learning for Efficient Scheduling of Ground-based Astronomical Observations
The Astronomical Journal
Ground-based astronomy
Observational astronomy
Astronomical techniques
Neural networks
Optical observation
title Deep Reinforcement Learning for Efficient Scheduling of Ground-based Astronomical Observations
title_full Deep Reinforcement Learning for Efficient Scheduling of Ground-based Astronomical Observations
title_fullStr Deep Reinforcement Learning for Efficient Scheduling of Ground-based Astronomical Observations
title_full_unstemmed Deep Reinforcement Learning for Efficient Scheduling of Ground-based Astronomical Observations
title_short Deep Reinforcement Learning for Efficient Scheduling of Ground-based Astronomical Observations
title_sort deep reinforcement learning for efficient scheduling of ground based astronomical observations
topic Ground-based astronomy
Observational astronomy
Astronomical techniques
Neural networks
Optical observation
url https://doi.org/10.3847/1538-3881/ade3dc
work_keys_str_mv AT haicao deepreinforcementlearningforefficientschedulingofgroundbasedastronomicalobservations
AT shaominghu deepreinforcementlearningforefficientschedulingofgroundbasedastronomicalobservations
AT junjudu deepreinforcementlearningforefficientschedulingofgroundbasedastronomicalobservations
AT xuchen deepreinforcementlearningforefficientschedulingofgroundbasedastronomicalobservations
AT shuqiliu deepreinforcementlearningforefficientschedulingofgroundbasedastronomicalobservations
AT shuaifeng deepreinforcementlearningforefficientschedulingofgroundbasedastronomicalobservations
AT bozhang deepreinforcementlearningforefficientschedulingofgroundbasedastronomicalobservations
AT yuchenjiang deepreinforcementlearningforefficientschedulingofgroundbasedastronomicalobservations