The Low-loss And Resource Reserved Time Deterministic Cyclic Scheduling

The operation system is responsible for the control and management of the computer hardware and software resources,and it reasonably organizes program operation and sharing of various resources.These tasks are processed and distributed by scheduling algorithms in the operating system to enable them to be executed in an orderly manner.In the scheduling process,the scheduling can be divided into real-time scheduling and non-real-time scheduling by different real-time processing objects.Real-time scheduling mainly schedules real-time tasks with deadlines,while non-real-time scheduling guarantees CPU utilization and throughput.As application requirements increase,the original non-real-time tasks face the need for low latency and fast response.Such an upgrade makes non-real-time scheduling unable to meet its needs,and using real-time scheduling will be difficult to meet its high utilization requirements.In solving such new soft real-time periodic tasks,so how to balance the factors such as delay,throughput,interrupt response and channel utilization is of great significance for the study of scheduling algorithms.This thesis proposes a new non-real-time operation system scheduling algorithm called Resource Reserved Time Deterministic Cyclic Scheduling(RRTDCS).This is an offline cache pre-allocated schedule.This scheduling method is a hybrid scheduling algorithm that combines time reservation and priority policies.The algorithm can be modified algorithmically for different task sets,and can be used in general methods or methods specific to high voltage conditions.And it can adapt to the current scheduling environment to optimize its own scheduling process,so as to ensure the certainty of the task scheduling law while dynamically allocating the time interval.Compared with the classic cyclic scheduling execution method,RRTDCS is more flexible.And its periodic looping strategy reduces the extra time overhead caused by task switching or preemption during task scheduling.This thesis uses Matlab for simulation,establishes RRTDCS scheduling model,and tests it,and simulates the current mainstream real-time scheduling schemes for comparison at the same time.This paper mainly tests for the completeness of delay and periodic tasks.The experimental results show that RRTDCS has better real-time performance and can break through the scheduling critical condition(?0.78)of the classic three-task set scheduling model,so that the schedulability of the task set including tasks can break through the sum of CPU utilization,that is (?).And it can adapt different test task sets with different expressions and scheduling models,thus achieving a balance of delay and completeness,and achieving better service quality.