Method to Evaluate Earliest Release Time of New Real-time Tasks
摘要
In a periodic real-time system scheduled with the Earliest Deadline First (EDF) algorithm,it is necessary to compress some current tasks to avoid overloading if new task requests to run.Compressing a task means that its period is prolonged while its computation time keeps unchanged.An interesting problem is to find the earliest time to release new tasks without any deadline missing,that is,the earliest smooth insertion time.In this paper,a general frame to calculate the earliest time with multiple rounds of deadline checking is given,which shows that the checking can be done from the request time of the new tasks.A smart way is provided and proved,which takes the value of the Δ checking of the current round as the time step to the next.These techniques potentially reduce the amount of the calculation and the number of the rounds of the checking to get the earliest time.Simulation results are also given to support the conclusion.