Research On Resource Assignment Models And Algorithms In Networks

Over the past decades,the Internet has been developed rapidly,and there are various kinds of businesses that rely on the Internet.Therefore,designing reasonable scheduling strategy for online businesses and allocating network resources appropriately are of great significance for improving user experience and system efficiency.The divisible load scheduling problem in distributed environment is an approximate description of many practical problems.Virtual Optical Networks(VONs)are considered to be the development trend of the next generation network.On this basis,combined with the actual situation,this article establishes two resource allocation models under the network environment,and efficient algorithms are proposed for them.The major achievements are as follows:1.At present,most of the researches on the divisible load scheduling are aimed at minimizing the makespan of the task.However,with the continuous improvement of the hardware performance,we can meet more needs of users while completing the tasks more rapidly,such as the computer's service quality,data precision and so on.Therefore,based on the existing divisible load scheduling model,this article proposed the concept of quality of service.A multi-objective optimization model considering the quality of service is established in the star network and aims to maximize the computer's service quality with the common optimization goal.When solving the multi-objective model of the divisible load scheduling considering the quality of service,a two population genetic algorithm with the framework of MOEA/D is put forward,two new crossover operators and an efficient mutation operator are designed to speed up the convergence of the solution and improve the accuracy of them.In order to prove the validity of the proposed model and algorithm,a series of experiments are carried out,and the results meet the goals in the model.2.The research of resource allocation for static traffic in VONs can be regarded as a sort of network planning.However,with the increasing complexity of the network platform,the issue of request security has become non-negligible.For some connection requests,they are forbidden specific physical nodes,such as: general request should avoid dedicated nodes.Based on the above,the concept of privileged node is proposed,and a resource allocation model considering the privileged nodes in VON has been established,the optimization goal is to minimize the index of used frequency slots.For solving the resource allocation problem in VONs considering privileged nodes,the existing adaptive algorithm are applied by the node mapping stage.Then a new routing algorithm is designed to find the candidate path set,which can avoid the failure of KSP(K-Shortest Path)algorithm in the case of privileged nodes.An improved genetic algorithm is adopted in the routing stage,which includes a cross operator based on uniform design,a mutation operator that effectively expands the search range of the algorithm,and dynamic rate are employed by the two operators to adjust the diversity of the population;the spectrum assignment stage uses the first fit strategy.The experimental results show that the proposed model is feasible and effective compared with the existing algorithm.