基于多目标数学规划的网络多源数据调度仿真

传统的目标网络多源数据调度方法通常以时间或费用为单一调度优化目标,无法实现任务完成时间以及任务执行成本之间的均衡,造成系统资源利用率较低。针对上述问题,提出一种基于多目标数学规划的网络多源数据调度方法。使用DAG构建网络多源数据流,确定多源数据调度任务模型的信任关系,以任务完成时间、任务完成成本、资源利用率为优化目标,建立多目标调度任务模型。对模型进行求解,在遗传算法变异操作中加入粒子群算法,对数据变异的方向与幅度进行调整,完成网络多源数据调度。仿真证明,所提方法相较于传统方法,在多源数据的调度上成本更低、资源利用率更高,并且调度任务目标完成时间更短。

Simulation of Network Multi-source Data Scheduling Based on Multi-Goal Mathematical Planning

Traditional multi-source data scheduling methods of target network take the time or cost as a single scheduling optimization goal,which cannot achieve the balance between task completion time and task execution cost,resulting in low utilization rate of system resource.Therefore,a network multi-source data scheduling method based on multi-objective mathematical programming was presented.Firstly,DAG was used to construct the network multi-source data flow and determine the trust relationship of multi-source data scheduling task model.Secondly,the multi-objective scheduling task model was built by taking task completion time,task completion cost and resource utiliza-tion as optimization objectives.After that,the model was solved,and then the particle swarm optimization(PSO)was introduced into the genetic algorithm mutation operation,so as to adjust the direction and amplitude of data varia-tion.Finally,the network multi-source data scheduling was completed.Simulation results prove that the proposed method has lower cost and higher resource utilization during the multi-source data scheduling than the traditional methods.Meanwhile,it has shorter time to complete the scheduling task.