| Space is a new research and activities area for human following ground and sea, it’s animportant place in some key areas especially national defense. With the development ofscience and technology, people’s requirement on high bandwidth and quality ofcommunication, single terrestrial fiber network cannot meet the demand of people. Withthe improvement of Inter-Satellite, Satellite-Ground, Space-Ground link performance,we can combine satellite network with mature terrestrial network to best serve thecommunication needs of users. Satellite-earth integrated network composed of two parts:constellation network and terrestrial network. In this integrated network, it’s importantfor designer to consider the following factors: physical topology, link capacitydimensioning, task scheduling and routing algorithm. We will researched capacitydimensioning and task scheduling problem that related to resource management in thispaper.For capacity dimensioning problem, we first presented a integrated satellite andterrestrial network model, and modeling the link dimensioning problem. On capacitydesign, we propose a unified mathematical framework using a two-stage stochasticprogramming formulation. The objective of the formulation is to find optimal linkcapacities that minimize the effective system cost. Stochastic programming has considerthe uncertainty of future flow, and give result of optimal link capacity and optimalrouting strategy. In addition, the effect of traffic distribution on capacity designing isstudied in two kinds of network traffic. We also study worst case static dimension forgeneral GEO network. The research result can provide useful reference for capacitydimension in hybrid network.For task scheduling problem, we first analysis the new property of constellationnetwork scheduling, then described the task scheduling problem and analyses theconstraints of it, and proposed a communication scheduling model in which theISLs(Inter Satellite Links) is built. Furthermore, we analyzed three key problem of taskscheduling, on this basis we design the scheduling strategy for SaSo-SNRS andMaMo-SNRS. At last, we use STK combined with VC software to simulate the proposed algorithm. The experimental results show that the algorithm we proposed is anacceptable scheduling strategy. |