Research Initiation: Fluid Network and Its Applications in Scheduling Control

研究将调查在生产系统、计算机和通信网络中出现的排队网络的调度控制问题。尽管受到了广泛的关注，但对于一般结构化网络的调度控制还没有一个系统的方法。研究者应该使用一个简单的、高度聚合的模型，即流体模型来解决调度控制问题。所采取的观点是，简单的模型往往提供了对复杂问题的洞察力，解决实际问题的方法，并为研究更复杂的模型提供了基础。最近的研究表明，流体模型方法可以将调度控制问题简化为求解一组线性规划问题，并且在某些情况下，正确解释的解可以是排队系统的最优解。研究将分四个步骤进行：(1)用流体模型对排队系统建模；(2)求解流体模型的调度控制问题；(3)将解解释为原排队系统的控制规则；(4)通过仿真对解进行测试。

Research will investigate scheduling control problems of queueing networks that arise in production systems, and computer and communication networks. Despite much attention, a systematic approach to the scheduling control of general structured networks is not available. The investigator should solve the scheduling control problems by using a simple and highly aggregated model, namely, a fluid model. The viewpoint taken is that the simple model often provides the insight to the complicated problem, the solution to practical problems, and the basis for studying more complicated models. Recent work has shown that the fluid model approach can reduce a scheduling control problem to solving a set of linear programming problems, and in some cases the solution when properly interpreted can be optimal for the queueing system. Research will be performed in four steps: (1) modeling queueing system by fluid model, (2) solving the scheduling control problem of the fluid model, (3) interpreting the solution into the control rule for the original queueing system, and (4) testing the solution by simulation.