Research On Adaptive Real-time Scheduling Algorithm For NC System

With the rapid development of next-generation information technology,the global machine tool industry has entered an era of automation centered on digital manufacturing technology.The development of national equipment manufacturing industry depends on the product quality and technical level of the machine tool industry,which is an important part of the national fundamental construction and has momentous significance for promoting social development.As a product of the deep integration of information technology and advanced manufacturing technology,NC technology is the key element to implement the functions of high speed,high efficiency,high accuracy,and intelligent numerical control(NC)system.NC system is a typical application of real-time system.Real-time scheduling algorithm is the core to ensure the completion of the functions of real-time system under the constraints of deadline and energy consumption.With the diversification of processing requirements and the complexity of application scenarios,both the internal components and external environment of NC system become more and more difficult to describe accurately by mathematical model,which brings new challenges to the real-time scheduling algorithm.In order to achieve the optimal allocation of scheduling resources between tasks,it is necessary not only to fully consider the real-time requirements of scheduling mechanism in NC system,but also to comprehensively consider the uncertainty of external environment,system energy consumption,task priority constraints or data dependence and other issues.Adaptive real-time scheduling algorithm combines the control theory,queuing theory,artificial intelligence and other methods of adaptive system with scheduling algorithm organically.The scheduling model and performance evaluation system of real-time system are analyzed and constructed by control theory.In the whole system operation process,the scheduling parameters are dynamically adjusted by continuously observing scheduling information and combining with classical real-time scheduling rules,which effectively solves the problem of dynamic load change and unpredictable system scheduling.This thesis focuses on the task scheduling problem in NC system,analyzes the specific scheduling requirements in complex internal and external environments,and applies the adaptive control method to the classical real-time scheduling algorithm,obtains some adaptive real-time scheduling algorithms with higher performance in specific environments on the basis of inheriting the existing real-time scheduling algorithms.The specific research contents are as follows :Firstly,according to the requirements of soft real-time task scheduling with dynamic load changes,a real-time scheduling model based on resource reservation and feedback theory is proposed.The dynamic adjustment of scheduling-related parameters is implemented by establishing the state feedback control equation between system performance and error.Then,aiming at the requirement of hybrid task scheduling on shared platform with limited hardware resources,an adaptive hierarchical scheduling framework is established based on ABS algorithm to accomplish resource isolation between hard real-time tasks and soft real-time tasks.Secondly,considering a fact that the task's WCET of pessimistic estimation is usually much longer than the actual execution time,two hybrid tasks scheduling algorithm based on adaptive total bandwidth server and resource recovery strategy are proposed respectively.The schedulability of above two algorithms are analyzed,and the pseudocode descriptions of the specific process are presented.Thirdly,aiming at scheduling requirements of periodic dependent tasks and energy consumption,the scheduling strategy based on hybrid genetic algorithm is proposed.The specific steps of the algorithm are described.The effectiveness of the algorithm is verified by simulation experiments.