Waiting Time Causal Relationships and Delay Cascades in Queueing Networks: A Data-Driven Approach
A queueing network refers to a special type of network that consists of several interconnected queues.The waiting time dynamics of queues are not only affected by supplier/demander impact factors, but also influenced by the interaction of related queues.In this paper, we propose a data-driven approach to extensively study the waiting time causal relationship between related nodes and its underline reasons.Specifically, we firstly utilize Structure Equation Modeling to identify whether two related queues have a waiting time causal relationship.Based on empirical data about cardiac care in Ontario, Canada, our experiment shows that the waiting time of related nodes do have a significant causal relationship.Then we develop a series Markovian queueing network model to examine whether and how the delay cascade (i.e., a small delay in one place resulting in delays elsewhere) leads to form such kind of causal relationship.Our simulation results show that the delays in one queue will spread to other related queues, so that the waiting time causal relationship are formed during this process.
Liang Luo Li Tao Zili Zhang
Faculty of Computer and Information Science, Southwest University
国际会议
昆明
英文
327-341
2014-05-01（万方平台首次上网日期，不代表论文的发表时间）