基于排队论的Modelica云仿真任务调度系统规划研究

Research on Modelica Cloud Simulation Resource Scheduling System Planning Based on Queuing Theory

为了提高Modelica云仿真系统任务调度的均衡性和效率,同时最大化资源提供者的利益,构建基于排队理论的云仿真技术任务调度系统模型,考虑等待队列长度和仿真系统的模块分配,结合排队理论分析任务调度策略,提出Modelica云仿真系统模块化的M/M/c串联排队模型方案。实验结果表明,与传统的单一M/M/c模型相比,M/M/c串联排队模型在相同成本指标下更能满足资源提供者对于系统效率的需求。

In order to improve the balance and efficiency of task scheduling of Modelica cloud simulation system and maximize the interests of resource providers, a cloud simulation technology task scheduling system model based on queuing theory is constructed, considering the waiting queue length and module allocation of the simulation system, the task scheduling strategy is analyzed in combination with queuing theory, a modular M/M/c series queuing model scheme for Modelica cloud simulation system is proposed. The experiment results show that: Compared with the traditional single M/M/c model, the M/M/c series queuing model can better meet the needs of resource providers for system efficiency under the same cost index.