Research On Resource Allocation For Space-based Information Network Based On Resource Reliability

The Space-Based Information Network(SBIN) is the key element in gaining battle situation in future information wars, so safely and reliably operating of this network is critical in obtaining the battle priority as well as acquiring the final victory of wars. Many factors can influence the safely and reliably operating capability of SBIN, among which resource reliability is a primary one. This thesis is dedicated to analyzing the impact of resource reliability on SBIN, as well as designing a resource allocating method to improve the reliability of the network based on resource reliability. Main contents of this thesis are as follows:Firstly, the problem background of the thesis is introduced, general resource allocating methods for SBIN based on resource reliability are analyzed, and correlative researches on the resource allocating problem for SBIN are presented.Secondly, based on possible exist problems of resource reliability for SBIN are desceibed,the impact of resource reliability on resource allocation for the SBIN is analyzed, and the resource allocating process of SBIN based on resource reliability is described. Primary constraints and basic hypotheses of the resource allocating process are presented, too. Then the calculating method for resource reliability of SBIN is analyzed, which is used as the criterion in resource selection during the resource allocating process.Thirdly, traits of tasks in resource allocation for SBIN are investigated, and a mathematical model based on the Constraint Satisfactory Problem (CSP) model is founded to formulate the resource allocating problem for SBIN based on resource reliability. The complexity of that model is then analyzed, and to deal with the high complexity, an improved genetic algorithm based on a local solution space escaping mechanism is presented to solve the model reliably and efficiently.Finally, a set of resource allocating prototype system for SBIN is brought forward on the basis of model and algorithm analysis. An instance is ultimately designed to illustrate the application mode and means of the method presented in this thesis on the actual SBIN.