Research And Realization On Adaptive Scheduling For Open Real-Time Systems

In view of open environment, hybrid tasks exist and the environments of real-time systems become more and more dynamic and unpredictable. This paper presents an adaptive scheduling algorithm. The algorithm is that the system detects changes of environmental constantly, than control system's real-time dispatch part to adjust according to the message of feedback using the scheduling strategy based on closed loop and the feedback. In concrete design, the hard real-time tasks and soft real-time tasks use different feedback control strategy. Hard real-time task Adaptive adjustment executes on the test-run stage, system distributes these priority according to deadline. After hard real-time tasks completing, system judges that whether the hard real-time tasks servers need to be adjusted and calculate the rate of adjustment. Soft real-time adaptive adjustments executes on the operational stage. System calculates the rate of feedback to the dispatcher and adjusts the soft real time system's server rate. At the same time, adjusting priority and the operation of request control soft real-time tasks as a supplementary.Finally, this paper verifies the adaptive algorithm base onμC/OS- II. Experimental result indicates that this algorithm may execute on the actual operation system. Hybrid type tasks can be received, and different type tasks scheduling according to different strategy. Server rate and the priority of tasks can adjust by environment regulation inside and outside.Adaptive scheduling is an effective measure to solve the open environment. Through adaptive adjustment, system distributes the resources reasonably, causes the different type tasks schedule coordinately. When the new tasks join, the system can carry on simply test of accept. Scalability and scheduling are enhanced and reliable safeguard of time restraint for the real-time application time is provided.