The Realization Of Multiprocessor Scheduling Algorithm And Simulation Framework Research

The Real-time operating system is widely used in the Embedded Development, recently. For the feature of that it can reasonable schedule, make overall plans and take all factors into consideration for different task with different important level, the key of the Real-time operation system is the schedule algorithm. The result of the task schedule research is helpful for the key system area in our country. This paper which is based on the theory of the task schedule of the Real-time operating system, researching and analyzing the common Real-time operating system,and the new improved Real-time scheduling algorithm and simulation framework are built. At last, the developed algorithm in the enterprise raw materials examination process has better effect.The study in this paper focus on the following areas:Firstly, this paper introduces several common Real-time operation systems currently: RTLinux, QNX, LynxOS, RTEMS, VxWorks. The main work is that introducing the way of the Real-time kernel implement and the Real-time performance optimization based on the such framework. At the base of that, analyzing the existing test method and combining of the real-time schedule algorithm theory, a new real-time schedule simulation framework-PSch based on high-level colored time Petri net is presented. The simulation framework not only solutions to the disadvantages of traditional Petri net, but also solutions to the disadvantages such as incomplete test parameter, weak capability of real-time supervision and generality. Experimental result shows that PSch can simulate and test the real-time schedule algorithm well.Taking the RTLinux for example, based on analyzing the kernel code, we research the way of the Real-time schedule implement of the kernel. Based on that, according to the multiprocessor feature, a new dynamic parallel scheduling algorithm of real-time multiprocessor, named Split-Parallel (SPara), is presented. The algorithm solved the problem that the previous algorithms such as Myopic, EDPF, etc, only judging by the deadline to schedule the tasks ,and the algorithm is developed to add the restrictive of the urgency, and an effective method as the task with long execution time. Furthermore, the multiprocessor scheduling algorithm which combine the theory of high-level colored time Petri net, is analyzed by modeling, and according to the model, an example of SPara algorithm was simulated and tested. The experimental result shows that the new multiprocessor scheduling algorithm-SPara is much better than the other algorithms in the aspect of the use ratio and scheduling success ratio. At last, the developed scheduling algorithm SPara put use in the raw-material-inspect system. The result shows that the new algorithm improve the inspect success ratio and checker's efficiency.