Deep Reinforcement Learning with Local Attention for Single Agile Optical Satellite Scheduling Problem
Open Access
Abstract
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.Keywords
Funding Information
- Ministry’s and National Commission’s Scientific Research Project (TJ2022B020200-003-020)
This publication has 41 references indexed in Scilit:
- Classification of Metal Handwritten Digits Based on Microwave Diffractive Deep Neural NetworkAdvanced Optical Materials, 2023
- Multi-satellite scheduling problem with marginal decreasing imaging duration: An improved adaptive ant colony algorithmComputers & Industrial Engineering, 2022
- Dynamic regional splitting planning of remote sensing satellite swarm using parallel genetic PSO algorithmActa Astronautica, 2022
- A competitive Markov decision process model and a recursive reinforcement-learning algorithm for fairness scheduling of agile satellitesComputers & Industrial Engineering, 2022
- Deep reinforcement learning and parameter transfer based approach for the multi-objective agile earth observation satellite scheduling problemApplied Soft Computing, 2021
- A Multi-objective Memetic Approach for Time-dependent Agile Earth Observation Satellite Scheduling ProblemComputers & Industrial Engineering, 2021
- Hierarchical scheduling for real-time agile satellite task scheduling in a dynamic environmentAdvances In Space Research, 2019
- An improved adaptive large neighborhood search algorithm for multiple agile satellites schedulingComputers & Operations Research, 2018
- Mission planning optimization of video satellite for ground multi-object staring imagingAdvances In Space Research, 2018
- An adaptive large neighborhood search metaheuristic for agile satellite scheduling with time-dependent transition timeComputers & Operations Research, 2017