Doctor of Philosophy
Statistics and Actuarial Sciences
This thesis consists of three contributing manuscripts related to waiting times with possible applications in health care. The first manuscript is inspired by a practical problem related to decision making in an emergency department (ED). As short-run predictions of ED censuses are particularly important for efficient allocation and management of ED resources we model ED changes and present estimations for short term (hourly) ED censuses at each time point. We present a Markov-chain based algorithm to make census predictions in near future.
Considering the variation in arrival pattern and service requirements, we apply and compare three models which best describe our data. We provide hourly predictions up to 24 hours in a day which will provide suggestions to ED managers on how to prevent over-crowding in their system. We illustrate our approach using 22 months data obtained from the ED of a hospital in south western Ontario.
The next two manuscripts extend the theory underlying the Accumulating Priority queues (APQs). We focus on the queues with two classes of customers and Poisson arrivals. The first work in this topic derives the stationary waiting time distributions for the class of lowest priority customers in an Affine Accumulating Priority queues (Affine APQs). APQs were first studied by Kleinrock (1964) and later revisited by Stanford et al (2014) where they obtained explicit solution for the Laplace Stieltjes Transform (LST) of the stationary waiting times for all classes of customers.
All subsequent publications on APQs, have assumed that all arriving customers accumulate priority credits over time starting from the same initial value (assumed, without loss of generality, to be 0). Whereas, our model studies Affine APQs which assume different initial priorities (without loss of generality in a two-class setting we assume the lowest class starts with 0 credit and the higher class customers with positive credit a. In this work we determine the waiting time distributions for the lower class of customers with Poisson arrivals and general service and present some numerical results for special cases of M/M/1, M/M/c and M/D/1. Inspired by health care applications, we have also considered a particular optimization problem related to the Affine APQ model, in order to select the optimum accumulation rate which allows for the lowest class customers to meet their associated KPIs.
We next focus on the Analysis of the Maximum priority processes in the context of Affine APQ. Maximum Priority Processes were first introduced in the context of APQs in Stanford et al (2014). We derive the LST of the stationary steady state distributions of the Maximum Priority Processes as recursive functions and derive the explicit solutions for the LSTs in classical APQ (i.e. a = 0). We employ this argument to present a new approach to determine the LST of waiting time distribution for an APQ with two-classes of customers under the M=M=1 discipline. Since the Analysis of the Maximum Priority Processes in this work is done for the general class of Affine APQs, it has provided the grounds for future researches to obtain the LST of the waiting time distributions in the Affine APQs.
Mojalal, Maryam, "Statistical Applications in Healthcare Systems" (2018). Electronic Thesis and Dissertation Repository. 5329.