Research And Implementation Of Resource Scheduling Algorithm Based On Machine Learning In IVCE Environment

Science and technology is in a rapid development of the times,while the development of electronic terminal equipment is the most prominent.Both the hardware and network conditions of the electronic terminal equipment are constantly meeting the daily needs of people,while they are showing excessive capacity at the same time.The consequence of this situation is that more and more equipment resources are left unused or wasted,causing a great waste.In order to solve this problem,the iVCE was born at the right time.This environment is built on the open Internet platform and it completes the management of heterogeneous and diverse resources by integrating the resources and mission requirement.The choice of resource scheduling strategy not only determines the efficiency of the operation of the virtual computing environment,but also affect the efficiency and service level of resource.It is one of the most important research topic in this environment.In this paper,we study the resource scheduling problem based on the real iVCE.In this paper,we first analyze the problems of existing resource scheduling algorithm in iVCE platform and put forward a simple"classification" scheduling method based on the characteristics of resources.Then a resource management system with this function is implemented in the actual iVCE.Finally,we proposed a improved solution based on machine learning algorithm aiming at the shortcomings of the simple classification method.The main work of this paper includes:(1)The virtual computing environment and its existing resource scheduling algorithm are studied deeply in this paper.We found that most of the existing algorithm put their emphasis on how to shorten the implementation time of task.These methods not only ignore the characteristics of resources but also cause some blind scheduling problems.To solve these problems,a thought which is based on a classification of resources is proposed in this paper and its applicability is analyzed as well.(2)A resource management system is implemented in this paper.Besides the basic functions,this system can also complete the simple classification towards resources and schedule by their types.(3)The functional classification criteria in the existing system is not rigorous and its results are not accurate.On the other hand,the data amount int iVCE is very huge.As a result,we consider using some clustering algorithm in machine learning area to improve the existing system.An improved FCM clustering algorithm based on genetic algorithm is proposed and the correctness and usability of the solution are analyzed by experiments.