Search

We investigate a general solution based on reinforcement learning for the agile satellite scheduling problem. The core idea of this method is to determine a value function for evaluating the long-term benefit under a certain state by training from experiences, and then apply this value function to guide decisions in unknown situations. First, the process of agile satellite scheduling is modeled as a finite Markov decision process with continuous state space and discrete action space. Two subproblems of the agile Earth observation satellite scheduling problem, i.e., the sequencing problem and the timing problem are solved by the part of the agent and the environment in the model, respectively. A satisfactory solution of the timing problem can be quickly produced by a constructive heuristic algorithm. The objective function of this problem is to maximize the total reward of the entire scheduling process. Based on the above design, we demonstrate that Q-network has advantages in fitting the long-term benefit of such problems. After that, we train the Q-network by Q-learning. The experimental results show that the trained Q-network performs efficiently to cope with unknown data, and can generate high total profit in a short time. The method has good scalability and can be applied to different types of satellite scheduling problems by customizing only the constraints checking process and reward signals. [ABSTRACT FROM AUTHOR]

Abstract The effective scheduling of multiple agile earth observation satellites plays a key role in improving the efficiency of the agile satellite observation system. Due to the complex constraints and large solution space (increases exponentially with the problem size), it has been a big challenge to effectively solve the multiple agile satellites scheduling problem. In this study, an edge computing framework is proposed in order to solve this problem in a flexible manner. In this framework, a central node and several edge nodes are included. The central node corresponds to the operation center of the satellite observation system while each edge node corresponds to an agile satellite. The central node filters tasks for edge nodes according to the time of scheduling period and visible time windows. Tasks in each edge node are scheduled according to the order of scheduling periods by a constructive heuristic algorithm based on the density of residual tasks (HADRT). The efficiency of this algorithm is proved. The time and space complexity are analyzed in detail. The proposed method is compared with other three advanced methods, and the experimental results show that HADRT could achieve higher task completion rate and reward rate than other algorithms with acceptable computing time. [ABSTRACT FROM AUTHOR]

The satellite image downlink scheduling problem (SIDSP) is included in satellite mission planning as an important part. A customer demand is finished only if the corresponding images are eventually downloaded. Due to the growing customer demands and the limited ground resources, SIDSP is an oversubscribed scheduling problem. In this paper, we investigate SIDSP with the case study of China’s commercial remote sensing satellite constellation (SuperView-1) and exploit the serial scheduling scheme for solving it. The idea is first determining a permutation of the downlink requests and then producing a schedule from the given ordered requests. A schedule generation algorithm (SGA) is proposed to assign the downlink time window for each scheduled request according to a given request permutation. A hybrid genetic algorithm (HGA) combined with neighborhood search is proposed to optimize the downlink request permutation with the purpose of maximizing the utility function. Experimental results on six groups of instances with different density demonstrate the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]

The autonomicity of self-scheduling on Earth observation satellite and the increasing scale of satellite network attract much attention from researchers in the last decades. In reality, the limited onboard computational resource presents challenge for the online scheduling algorithm. This study considered online scheduling problem for a single autonomous Earth observation satellite within satellite network environment. It especially addressed that the urgent tasks arrive stochastically during the scheduling horizon. We described the problem and proposed a hybrid online scheduling mechanism with revision and progressive techniques to solve this problem. The mechanism includes two decision policies, a when-to-schedule policy combining periodic scheduling and critical cumulative number-based event-driven rescheduling, and a how-to-schedule policy combining progressive and revision approaches to accommodate two categories of task: normal tasks and urgent tasks. Thus, we developed two heuristic (re)scheduling algorithms and compared them with other generally used techniques. Computational experiments indicated that the into-scheduling percentage of urgent tasks in the proposed mechanism is much higher than that in periodic scheduling mechanism, and the specific performance is highly dependent on some mechanism-relevant and task-relevant factors. For the online scheduling, the modified weighted shortest imaging time first and dynamic profit system benefit heuristics outperformed the others on total profit and the percentage of successfully scheduled urgent tasks. [ABSTRACT FROM AUTHOR]

With the current miniaturization and cost reduction trend, many small imaging satellites are deployed in low Earth orbit (LEO). Efficient scheduling of transmission links is crucial for handling the large volume of imaging data in the time-varying inter-satellite broadband communication networks. This ensures optimal utilization of inter-satellite link resources and gives rise to a novel problem known as the L arge-volume LEO satellite I maging D ata N etworked T ransmission S cheduling P roblem (LLSIDNTSP). This problem requires the integrity of the imaging data received on the ground station. A S pecific T ime- E volving G raph (STEG) model is introduced to formulate this problem, while a highly effective S equential T wo- P hased H euristic A lgorithm (STPHA) is proposed to solve it. STPHA determines an optimized C ontact P lan (CP) with a C ontact- F ast C onstruction A lgorithm (CFCA) in the first phase. With this CP as input, it produces an optimal transmission schedule with Linear Programming in the second phase. The CFCA incorporates a novel heuristic called M ission R esidual V olume F actor(MRVF), which guides the selection of a proper inter-satellite and satellite-to-ground link at each construction step. Extensive simulation results demonstrate the effectiveness of the proposed STPHA. In particular, it can achieve the optimal solution for small-sized scenarios with one-hundredth of the computing time used by the exact solver CPLEX. Also, it is consistently better than other heuristic algorithms adapted from the literature for large-sized scenarios. Additional experimental analysis is conducted to showcase the effectiveness of the innovative components of STPHA. • We study a new satellite imaging data transmission scheduling problem (LLSIDNTSP). • A Specific Time-Evolving Graph (STEG) model is proposed to formulate the LLSIDNTSP. • A Sequential Two-Phased Heuristic Algorithm(STPHA) is proposed to solve the problem. • Extensive experimental results demonstrate high efficiency of the proposed method. [ABSTRACT FROM AUTHOR]

In this paper, a new problem called stochastic quadratic multiple knapsack problem (SQMKP) is studied. SQMKP is the extension of the quadratic multiple knapsack problem (QMKP) with stochastic factors. Different from QMKP, the objective of SQMKP is no longer to find a single global optimal solution, but to find a solution with the best expected quality under all possible cases. We use the recoverable robustness technique to tackle these stochastic factors. The idea of the recoverable robustness is to first find a solution that is feasible for the static QMKP, and then adjust the solution through a simple recovery algorithm according to the emerging stochastic scenarios. Following the idea of the recoverable robustness technique, we propose a repair-based approach (RBOA) which involves three key components – an instance generator that is able to create a set of representative scenarios, a static QMKP algorithm that is able to provide a set of potentially robust solutions and a simple fast repair algorithm that can quickly restore the feasibility of a solution with respect to a given stochastic scenario. Experimental results on a set of stochastic instances extended from 90 well-known QMKP benchmarks demonstrate the efficacy of the proposed RBOA. [ABSTRACT FROM AUTHOR]

The unconstrained binary quadratic programming (UBQP) problem is a difficult combinatorial optimization problem that has been intensively studied in the past decades. Due to its NP-hardness, many heuristic algorithms have been developed for the solution of the UBQP. These algorithms are usually problem-tailored, which lack generality and scalability. To address these issues, a heuristic algorithm based on deep reinforcement learning (DRLH) is proposed in this paper. It features in inputting specific features and using a neural network model called NN to guild the selection of variable at each solution construction step. Also, to improve the algorithm speed and efficiency, two algorithm variants named simplified DRLH (DRLS) and DRLS with hill climbing (DRLS-HC) are developed as well. These three algorithms are examined through extensive experiments in comparison with famous heuristic algorithms from the literature. Experimental results show that the DRLH, DRLS, and DRLS-HC outperform their competitors in terms of both solution quality and computational efficiency. Precisely, the DRLH achieves the best-quality results, while DRLS offers a high-quality solution in a very short time. By adding a hill-climbing procedure to DRLS, the resulting DRLS-HC algorithm is able to obtain almost the same quality result as DRLH with however 5 times less computing time on average. We conducted additional experiments on large-scale instances and various data distributions to verify the generality and scalability of the proposed algorithms, and the results on benchmark instances indicate the ability of the algorithms to be applied to practical problems. [ABSTRACT FROM AUTHOR]