Research On Cyber-Physical System For Task Scheduling Algorithm

Cyber Physical System (CPS) is a new intelligent distributed system where information computation and physical processes are closely combined and collaborated. It is a new generation of technology products after the cloud computing and Internet of things technology. Centered on information, CPS integrates computing, communications and control and ultimately provides people with flexible, credible and efficient service.Compared with the traditional distributed systems, CPS pays more attention to the task optimization scheduling and reasonable allocation, in order to achieve real-time sensing and dynamic control for large-scale complex engineering system, and provides flexible and efficient service for users. However, the complexity of the physical environment, strong heterogeneity of resources, dynamic topology of network structure has brought the huge challenge to task scheduling in CPS. As traditional scheduling algorithms cannot meet the demand of the overall performance of CPS, the study is carried out from sensing and computing task of CPS respectively, regarding its complex tasks, a reasonable scheduling algorithm was proposed to realize efficient organization and allocation of dynamic resource, the overall performance of CPS is improved.The main contents of dissertation are:(1) First we make a comprehensive analysis of the basic characteristics and architecture about CPS, and we introduce the related technology about task scheduling and traditional task scheduling algorithm to prepare for the later research content.(2) According to complex sensing task scheduling problem of CPS,a multi-objective optimization scheduling algorithm based on improved particle swarm is proposed in this paper. This scheduling algorithm retains the characteristic of fast convergence speed and high efficiency of particle swarm algorithm. And it combines adaptive levy flight strategy to improve global optimization capability of the algorithm. On the precondition of minimum execution time of sensor task, the algorithm effectively realizes load balancing for sensor nodes and prolong the network life cycle.(3) In order to meet Multiple QoS target demand for task of users, according to complex calculation task scheduling problems of CPS, this paper proposes a hybrid scheduling algorithm based on multi-dimension QoS. The proposed method integrates the advantages of good robustness and high efficiency of Artificial Bee Colony, fast global search capability of SA, and it avoids the defects of artificial colony algorithm which is easy to fall into local optimum. The simulation results show the proposed algorithm not only meets the target requirements of multiple QoS for task of users, but also can effectively improve the efficiency of mapping between tasks and resources.Finally, we make a simulation and performance analysis for the above two algorithms respectively. The experimental results show that compared with the traditional algorithm, the proposed algorithm has better performance in task execution time and energy consumption, it can meet users’ multiple QoS target demand for task. The two algorithms is feasible.