Research On Highest Value-rate First Scheduling Algorithm For Real-time System Under Overload Conditions

With the rapid development of social economy, real-time system was widely used in areas such as aerospace and industrial control. The real-time scheduling algorithm is the core of real-time systems and plays a crucial role in the real-time system. The reasonable real-time system scheduling algorithm has a great significance of improving the performance of the system and ensuring real-time task scheduling.This paper analyzes the current and common real-time system scheduling algorithm, such as the earliest deadline first EDF algorithm, monotonous rate RM algorithm, the highest value most HVF priority algorithm, minimum density LL algorithm and highest value density first HVDF algorithm etc. Based on these algorithms, proposing the highest priority scheduling algorithm which is based on the value of a single processor system rate under the condition of overload system.This paper uses unified modeling quad Ti (Ai Ci Di Vi) to describe the task information. Ai represents the arriving time of the task,Ci represents the needed time of mission performance, D, represents the ending time of the task, Vi represents the value of the task. For a task which is dependent, using adjacency matrix represents dependencies between the tasks.According to the dependencies of the task, dividing the task into 2 parts: independent tasks and dependent tasks. According to the quality of real-time, dividing into 2 parts:the real-time tasks and non real-time tasks. And taking different approaches according to the difference between them, they are all under the time-constraint condition when comparing the value ratio, we can get the conclusion that the higher value rate, higher level priority of the task, and vice versa.Independent task scheduling algorithm adopts the idea of dynamic scheduling, allowing a higher-priority task preemption, task uses scheduling algorithm to operate during the running of task. Independent real-time and non real-time tasks do not have deadline, so all tasks meet the condition of constraint time, name that, finish the task before the deadline, we just need to compare task value rate of the scheduling algorithm, the higher value rate, the higher priority. For real-time tasks without dependencies, there is a real-time task deadline, task need to meet the time constraints condition, and then we need to compare value rate. When independent real-time tasks mixed with the non real-time tasks, all the real-time task priority is higher than that of non real-time tasks because of the deadline, and then we can use the scheduling algorithm to calculate.The dependent scheduling task adopts the idea of static scheduling. First of all, according to the task dependencies during the information management processing, on the premise of meeting the time constraints, determining priorities based on the dependency, dependent tasks groups determine the priority value rate through the process of comparing the value rate.