Optimization Strategy On The Schedulability Of DRT Task Set Based On Shaping Method

With the development of the Internet of Things such as Wearable Device,Telehealth Homecare and Wireless Smart Home,the real-time embedded system is becoming more popu-lar in our daily life than before.Real-time task graphs are used to describe complex real-time systems.And DRT(Digraph Real-Time)task model with high expressiveness plays an im-portant role among these real-time task models in recent years.It also becomes a hot issue in the field of real-time embedded system to improve the schedulability of DRT task model.An recent research shows that the workload of the systems are typically bursty,which makes their schedulability difficult to be met even with sufficient resource in the long term.In this thesis,we propose to use shaping method to improve the system scheulability under a preemptive uniprocessor using the static priority.The idea in this thesis is to insert artificial delays to the release times of certain vertices of a task graph to get a new graph with a smoother workload than the graph befor transforming,which is easier to be scheduled the lower tasks,while still meeting the timing constraints of the original task graph.Delaying the release time of a vertex may smoothen the workload of some paths of thetask graph,but at the same time make the workload of other paths more bursty.We devel-oped efficient techniques to search for an appropriate release time delay for each vertex with monotonic schedulability improvement guarantees.Experiments with randomly generated task systems show that the proposed optimizationstrategy based on shaping method can turn a significant number of task systems that was origi-nally unschedulable into schedulable,and the shaping procedure is very efficient and can easily handle realistic-size task graph systems in very short time.