Research On Hierarchical Multi-task Scheduling In The Distributed Command Systems

The carrier-based distributed command systems combine a large number of sensors, weapons, all kinds of communication equipments together, forming a information integration platform with powerful abilities of offense and defense. The distributed combat command system consists of many processors of different functions, and these processors receive a large number of different data collected by sensors instantaneously to make a further analysis of data and to publish the command, which challenges the combat command system to deal with sensor's data instantaneously and perform tasks reliably. The distributed command scheduling system is the core of the distributed command system. The design of task scheduling system model and task scheduling algorithm are the key technologies of distributed command scheduling system. Therefore, the research on task scheduling problems in distributed command system is especially urgent.The distributed task scheduling command is the research how to reasonably and reliably make an efficient use of processor resources. Its purposes is in order to maximize the use of the existing computer resources, while minimizing the communications between the processors and reducing the turnaround time for the tasks. Aiming at the characteristics of distributed combat command system covering a large scale network and multi-release task, the paper research the task scheduling algorithm on multi HTA (Hierarchical Task Allocation). By stratifying the scheduler, it makes the distributed command form a divide and rule structure, so as to solve the problem of shortening turnaround time of the scheduling tasks. Considering the distributed combat command system's loose coupling, this paper solves the task transmission problems by using publish and subscribe mechanism.The NS2 simulation results show that:the multi-layered task scheduling model, the scheduler's internal structure and the transport mechanism of publish and subscription are able to be combined effectively. The hierarchical scheduling algorithm has higher efficiency than the A* task scheduling algorithm and the cluster-based task allocation strategy. It makes the failure rate of task scheduling lower, decreases the weighted value of delay task, and significantly improves CPU utilization. With the scale of distributed command system increasing, the experiment showed that the algorithm exhibited the higher performance in the failure rate of task scheduling, the weighted value of delay task, and CPU utilization.