Resources Optimization Research For Tasks In Satellite Network

The satellite networks composed of various types satellite system on different orbit is the main information technology in the future. The resources optimization for task in the satellite network is assigned the related sub-tasks of multi-tasks to suitable satellite under the constraint condition, in order to make full use of satellite resources. As the development of modern satellite and distributed satellite system rapidly, and the research of satellite network growing, and the using of satellite network resources growing, the investigation on such problems is of great significance both from the theory or application viewpoints.We often to apply reconnaissance, computation and storage resources when the user's make a request for scout task. This thesis puts forward a resources optimization algorithm based on genetic algorithm, It can be given reasonable allocation scheme of resources and the execution sequence of tasks.Firstly, this thesis analyzes the management architecture of satellite network and optimization algorithm, studied the basic process and operating methods of genetic algorithm. Then give the optimization model and the flow of the resources optimization. And we designed the optimized algorithm based on genetic algorithm for satellite network. In order to express the task and sub-task more expediently, we designed three two-dimensional matrix: resource occupy matrix, sub-task poset matrix and time consumption matrix. In the operation of genetic algorithm, we adopted the decimal encoding method, put the completion time shortest of all task as the design goal of the fitness function. And designs a new subtask-based order crossover(STOX), which can inherit the sub-task order of father generation, guarantee progeny chromosome is feasible solutions, and the operate is very simple. Then we designed and realized the simulation experiment system based on the optimization algorithm.Finally, this thesis gives an example and solves it by the prototype system, and analyzed the solutions. The simulation results imply that the generic algorithm can obtain the approximate solution of the optimization problem in a shorter time.