Research On The Dynamic Adaptive Scheduling Algorithm Of Tasks In Real-time Embedded Systems

Remote temperature sensors,sports bracelets,smart phones,and remote sensing satellites,all kinds of real-time embedded devices have been quietly integrated into our lives.The reasons for the diversity of embedded devices are mainly determined by their working environment and work purpose.For example,decide the processing core is a single processing core or a multi-processing core,the processing mode within the processing core is isomorphic work mode or heterogeneous work mode depending on the complexity of the equipment application field,according to the equipment work environment to determine whether the supply of energy equipment is limited.The external working environment and the internal working environment of real-time embedded devices are always changing.If the device's built-in task scheduling and execution strategy can be changed according to the changes in the device's own work environment,it will undoubtedly allow the device to integrate into the work environment better,give full play to the full performance of the device,and make the device's own value more fully fulfilledThis paper first introduces the research background and development status of real-time systems,real-time tasks and task scheduling.After that,it describes the current classic task scheduling execution strategies.After the elaboration,the main work of this paper is introduced:(1)Analyze the idle time algorithm,deadline algorithm and value algorithm in the single processing core work environment.According to the deficiencies in the algorithm,this paper proposes a dynamic scheduling strategy for the real-time tasks and for the task energy consumption under the single processing core.Based on the classic scheduling strategy,the real-time demand task and the energy demand task are adapted so that the algorithm can find a balance between real-time performance and energy saving and make sure the equipment's capabilities can be more perfectly displayed.(2)Analyze the Min-Min strategy in the multi-processing core work environment and put forward the task real-time dynamic adaptive scheduling improvement strategy under the multi-processing core.The execution characteristics of Min-Min have been compatible and adapted so that the algorithm can handle two different types of tasks in a balanced way,which improves the real-time performance of the device.(3)Finally,a self-adaptive task scheduling strategy based on neural network model is proposed for complex real-time embedded device system with mixed tasks,so that the device can continuously analyze the current dominant task types and a variety of possible scheduling results during the scheduling process.Align different types of dominant tasks with scheduling parameters that have the best scheduling results,so that devices can adapt to hybrid tasks and improve device processing capabilities.