A New Architecture of Task Classification Oriented Atomization Computing Scheduling

Abstract

Due to the huge amount of data processed by the advanced large-scale intelligent manufacturing system, the computing mode of its data has also changed from node centered to task centered. This paper proposes a role object centered online mode flow scheduling method. Taking the user oriented distributed computing as the model, by establishing the task scheduling method of role mode flow, the cooperation behavior in the scheduling cooperation mode flow of large intelligent manufacturing nodes is combined with the behavior flow defined by the organization, so as to establish the role model behavior. Ensure that the processing of task role nodes in large intelligent manufacturing system is from the center to the edge, so that multi task role nodes can process related processes according to task patterns. This method can effectively solve the problems of large intelligent manufacturing system, such as variable tasks, complex human-computer interaction, slow node task import, high data model power consumption, frequent data interruption, etc.