Satellite mobile edge computing (SMEC) achieves efficient processing for space missions by deploying computing servers on low Earth orbit (LEO) satellites, which supplements a strong computing service for future satel...
详细信息
Satellite mobile edge computing (SMEC) achieves efficient processing for space missions by deploying computing servers on low Earth orbit (LEO) satellites, which supplements a strong computing service for future satellite-terrestrial integrated networks. However, considering the spatio-temporal constraints on large-scale LEO networks, inter-satellite cooperativecomputing is still challenging. In this paper, a multi-agent collaborative task offloading scheme for distributed SMEC is proposed. Facing the time-varying available satellites and service requirements, each autonomous satellite agent dynamically adjusts offloading decisions and resource allocations based on local observations. Furthermore, for evaluating the behavioral contribution of an agent to task completion, we adopt a deep reinforcement learning algorithm based on counterfactual multi-agent policy gradients (COMA) to optimize the strategy, which enables energy-efficient decisions satisfying the time and resource restrictions of SMEC. An actor-critic (AC) framework is effectively exploited to separately implement centralized training and distributed execution (CTDE) of the algorithm. We also redesign the actor structure by introducing an attention-based bidirectional long short-term memory network (Atten-BiLSTM) to explore the temporal characteristics of LEO networks. The simulation results show that the proposed scheme can effectively enable satellite autonomous collaborative computing in the distributed SMEC environment, and outperforms the benchmark algorithms.
To cope with the increasing information in sensor networks, distributed cooperative computing is used, to exchange and utilize the data. However, the data in each specific sensor network is heterogeneous with specific...
详细信息
ISBN:
(纸本)9781450353526
To cope with the increasing information in sensor networks, distributed cooperative computing is used, to exchange and utilize the data. However, the data in each specific sensor network is heterogeneous with specific formats and semantics. Thus, it is a huge conundrum to integrate, share and interoperate these sub sensor networks. To solve the above problem, a real-time novel distributedcooperative model for Isomerism Multi-Sensor network (DCIMS), is proposed in this paper. The DCIMS architecture is based on distributedcooperative data caching and forwarding, which can handle massive, semantic and real-time data stream at each Sensor Gateway node. A new filtering algorithm named Post-Tree-Automata filter (PTAfilter) is proposed based on unranked tree automata and suffix technologies. The proposed filter can process real-time XML streaming (which contains uncertain semantic elements), and understand more complex subscription structures. Both simulation results and theoretical analysis demonstrate that the proposed method improves the query speed, reduces redundant transmission intermediate states.
As the size and cost of embedded devices continue to decrease, it becomes economically feasible to densely deploy networks with very large quantities of such nodes, and thus enabling the implementation of networks wit...
详细信息
ISBN:
(纸本)9781479907984
As the size and cost of embedded devices continue to decrease, it becomes economically feasible to densely deploy networks with very large quantities of such nodes, and thus enabling the implementation of networks with increasingly larger number of nodes becomes a relevant problem. In this paper we describe a novel algorithm to obtain the number of live nodes with a very low time-complexity. In particular, we develop a mechanism to estimate the number of nodes or the number of proposed values (COUNT), with a time complexity that increases sublinearly with the number of nodes. The approach we propose is based on the wise exploitation of dominance-based protocols and offers excellent scalability properties for emerging applications in dense Cyber Physical Systems.
This paper introduces the concept of grid computing and its characteristics. Basing on analyzing the characteristics of grid computing, we apply grid computing to GIS field to construct grid GIS, which is the applicat...
详细信息
ISBN:
(纸本)9780819469144
This paper introduces the concept of grid computing and its characteristics. Basing on analyzing the characteristics of grid computing, we apply grid computing to GIS field to construct grid GIS, which is the application of grid computing on geographic information system with grid computing as the basic running environment. According to the architecture of grid and integrating the characteristics of GIS, we design the architecture of grid GIS, which has three layers. They are grid GIS resource layer, grid GIS middleware layer and grid GIS application layer. Among them, the grid GIS middleware layer is the most important. Then this paper expatiates on our research on grid GIS middleware, which is about the design and development of distributed cooperative computing GIS software. The key technologies of the distributed cooperative computing GIS software are discussed, which include technology of global spatial resource management, spatial data computing task allocation and management, system consistency mechanism and system security mechanism. The implementation process is also presented. At last, this paper presents the further researches of the distributed cooperative computing GIS software.
暂无评论