Research And Implementation Of Resource Coordination Scheduling Technology In Space-Ground Integrated Network

Currently,the space-ground integrated network is a hot spot in the development of global technology and industry.In the space-ground integrated network,the states of node are difficult to obtain because of the complex network topology,the far distance among nodes and the high delay.And there are multiple constraints on coordinated resource scheduling.In ground network,the coverage is limited,but the routing and computing capacity is sufficient and the network topology is stable.In space network,the coverage is wide,but nodes lack of computing capacity,the routing changes frequently,and nodes are dynamic and have limited visibility following time windows.Coordinating the resources and realizing efficient resource scheduling in the space-ground integrated network is one of the problems that is needed to be solved.To solve the above problems,the resource coordination scheduling technologies in the space-ground integrated network is focused in this thesis.The research mainly includes two aspects,the acquisition of node states and the resource scheduling algorithm.At first,a prediction based on node perception algorithm is proposed to predict the real-time node states by historical data.Two prediction models,ARMA and LSTM,which have similar fitting effects,are studied.Because ARMA requires more complicated calculations and LSTM has high real-time performance,the deep learning network based on LSTM is used for node state perception.A multi-layer neural network is added in front of the LSTM in the model,which effectively increases the depth of the entire model to make the prediction effect better.In view of the topological constraints of the space-ground integrated network,resource scheduling is studied,including the relay link and the space-ground link.And the resource scheduling algorithm is proposed,which includes the relay scheduling method based on the storage-and-forward for the relay link,the space-ground scheduling method based on Min-Min method for the space-ground link and the overall scheduling model based on the ant colony algorithm for the full link.Because the topology of the relay link is highly dynamic,the relay scheduling method based on store-and-forward uses data to be subcontracted,making full use of other idle node resources,which is a good solution to the problem of limited visibility by time windows.Because the topology of the space-ground link is fixed,the scheduling can be considered as multi-task node allocation on limited resources.The improved Min-Min algorithm is used for the space-ground scheduling,which solves the load through fuzzy load classification and local readjustment based on traditional Min-Min.The full link optimization considers the relay link and the ground-ground link as a whole,and uses the ant colony algorithm as the outer overall scheduling algorithm,which improves the traditional pheromone update rules of ant colony algorithm in order to prevent the algorithm from falling into a local optimum.The simulation system for resource coordination scheduling technology in the space-ground integrated network is designed and implemented.The key technologies proposed in this thesis are applied to the system.The system supports the initialization of the network based on node datas and the display of the network topology,supports the display of task deployment and log of the network.In the end,a series of functional tests and typical scenario tests for the system not only prove the effectiveness of the system,but also further illustrate the effectiveness of the technical research section.