A Queueing Network Model for Refugee Language Courses in Vancouver

Abstract

A surge of over 1,600 landed refugees in Vancouver has caused a spike in demand for English language courses resulting in longer than usual wait lists. With no accurate information regarding expected wait times, refugees often choose suboptimal waiting queues, contributing to the large variance observed in wait times. The purpose of this paper is to explore the bene1ts of a queueing network which assigns refugees to queues based on the expected wait times, with the intent of reducing total service times given the current resources. Furthermore, a range of priority policies are explored in order to observe their effect on the increasing refugee student population. By considering the distinct wait lists that form for each level of language course within each learning centre, a network of queues can be devised with each queue represented by a node. The nodes are organized by course level l, with each node in level l connected to each node in level l + 1 by a unique directed edge. For each queue in the network, the expected wait and service times are approximated using factors such as class capacity, current queue length, and departure rate. This queueing network solution reduces expected wait times, lowers the variance in wait times, and results in faster integration of refugees into Canadian communities and the labour force.