Deep Reinforcement Learning with Local Attention for Single Agile Optical Satellite Scheduling Problem

This paper investigates the single agile optical satellite scheduling problem, which has received increasing attention due to the rapid growth in earth observation requirements. Owing to the complicated constraints and considerable solution space of this problem, the conventional exact methods and h...

Full description

Saved in:
Bibliographic Details
Main Authors: Zheng Liu, Wei Xiong, Chi Han, Xiaolan Yu
Format: Article
Language:English
Published: MDPI AG 2024-10-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/24/19/6396
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850284277938257920
author Zheng Liu
Wei Xiong
Chi Han
Xiaolan Yu
author_facet Zheng Liu
Wei Xiong
Chi Han
Xiaolan Yu
author_sort Zheng Liu
collection DOAJ
description This paper investigates the single agile optical satellite scheduling problem, which has received increasing attention due to the rapid growth in earth observation requirements. Owing to the complicated constraints and considerable solution space of this problem, the conventional exact methods and heuristic methods, which are sensitive to the problem scale, demand high computational expenses. Thus, an efficient approach is demanded to solve this problem, and this paper proposes a deep reinforcement learning algorithm with a local attention mechanism. A mathematical model is first established to describe this problem, which considers a series of complex constraints and takes the profit ratio of completed tasks as the optimization objective. Then, a neural network framework with an encoder–decoder structure is adopted to generate high-quality solutions, and a local attention mechanism is designed to improve the generation of solutions. In addition, an adaptive learning rate strategy is proposed to guide the actor–critic training algorithm to dynamically adjust the learning rate in the training process to enhance the training effectiveness of the proposed network. Finally, extensive experiments verify that the proposed algorithm outperforms the comparison algorithms in terms of solution quality, generalization performance, and computation efficiency.
format Article
id doaj-art-3da67f0ed0f843f397c28adac85bb11e
institution OA Journals
issn 1424-8220
language English
publishDate 2024-10-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-3da67f0ed0f843f397c28adac85bb11e2025-08-20T01:47:37ZengMDPI AGSensors1424-82202024-10-012419639610.3390/s24196396Deep Reinforcement Learning with Local Attention for Single Agile Optical Satellite Scheduling ProblemZheng Liu0Wei Xiong1Chi Han2Xiaolan Yu3National Key Laboratory of Space Target Awareness, Space Engineering University, Beijing 101416, ChinaNational Key Laboratory of Space Target Awareness, Space Engineering University, Beijing 101416, ChinaNational Key Laboratory of Space Target Awareness, Space Engineering University, Beijing 101416, ChinaNational Key Laboratory of Space Target Awareness, Space Engineering University, Beijing 101416, ChinaThis paper investigates the single agile optical satellite scheduling problem, which has received increasing attention due to the rapid growth in earth observation requirements. Owing to the complicated constraints and considerable solution space of this problem, the conventional exact methods and heuristic methods, which are sensitive to the problem scale, demand high computational expenses. Thus, an efficient approach is demanded to solve this problem, and this paper proposes a deep reinforcement learning algorithm with a local attention mechanism. A mathematical model is first established to describe this problem, which considers a series of complex constraints and takes the profit ratio of completed tasks as the optimization objective. Then, a neural network framework with an encoder–decoder structure is adopted to generate high-quality solutions, and a local attention mechanism is designed to improve the generation of solutions. In addition, an adaptive learning rate strategy is proposed to guide the actor–critic training algorithm to dynamically adjust the learning rate in the training process to enhance the training effectiveness of the proposed network. Finally, extensive experiments verify that the proposed algorithm outperforms the comparison algorithms in terms of solution quality, generalization performance, and computation efficiency.https://www.mdpi.com/1424-8220/24/19/6396single agile optical satellite schedulingdeep reinforcement learninglocal attentionadaptive learning rate
spellingShingle Zheng Liu
Wei Xiong
Chi Han
Xiaolan Yu
Deep Reinforcement Learning with Local Attention for Single Agile Optical Satellite Scheduling Problem
Sensors
single agile optical satellite scheduling
deep reinforcement learning
local attention
adaptive learning rate
title Deep Reinforcement Learning with Local Attention for Single Agile Optical Satellite Scheduling Problem
title_full Deep Reinforcement Learning with Local Attention for Single Agile Optical Satellite Scheduling Problem
title_fullStr Deep Reinforcement Learning with Local Attention for Single Agile Optical Satellite Scheduling Problem
title_full_unstemmed Deep Reinforcement Learning with Local Attention for Single Agile Optical Satellite Scheduling Problem
title_short Deep Reinforcement Learning with Local Attention for Single Agile Optical Satellite Scheduling Problem
title_sort deep reinforcement learning with local attention for single agile optical satellite scheduling problem
topic single agile optical satellite scheduling
deep reinforcement learning
local attention
adaptive learning rate
url https://www.mdpi.com/1424-8220/24/19/6396
work_keys_str_mv AT zhengliu deepreinforcementlearningwithlocalattentionforsingleagileopticalsatelliteschedulingproblem
AT weixiong deepreinforcementlearningwithlocalattentionforsingleagileopticalsatelliteschedulingproblem
AT chihan deepreinforcementlearningwithlocalattentionforsingleagileopticalsatelliteschedulingproblem
AT xiaolanyu deepreinforcementlearningwithlocalattentionforsingleagileopticalsatelliteschedulingproblem