09h00 - 09h25
Modelling Patient Flow in an Emergency Department: a Petri-Net Approach
We present a modelling approach based on Petri nets for patient flow in an emergency department. This method was applied to the Hotel-Dieu hospital in Paris and allowed an accurate simulation of
waiting times. The simulation has given more insights on the bottlenecks of the system.
09h25 - 09h50
Accommodating Patients’ Preferences in Appointment Scheduling: A Multiagent System Model
The vast majority of appointment scheduling systems that accommodate patient preferences are not automated due to the challenges in designing effective mechanisms for eliciting patient preferences and solving the service time allocation models. We propose multiagent system models that accommodate patient preferences in dynamic healthcare scheduling environments. The models tackle patient preferences elicitation and service time allocation complexities using an iterative bidding framework.
09h50 - 10h15
Using Simulation to Improve the Efficiency of a Radio-Oncology Department
The admistrators of l’Hôpital Cité de la Santé in Laval are confronted with uncertain decisions on the functioning of operations to meet the standard imposed by the Ministry of Health and Human Services of Quebec. Through simulation, a validation of organizational resources, a standardisation of physician tasks and new rules for machine reservation are explored. The objective is to provide patients with the shortest waiting time, and maintain a quality level treatment.
10h15 - 10h40
Online Optimization of Radiotherapy Patient Scheduling
In the province of Quebec, the number of oncology treatment centers is growing in order to answer the increasing number of cancer cases and the aging population. An efficient radiotherapy patient scheduling on the linear accelerators (lineacs) is crucial. First, it should ensure the delivery of the treatment with respect to all deadlines. Second, it should efficiently use the material resources to increase the number of patients seen in the center on a period of time. Different objectives and constraints have to be taken into account. The number of delayed patients and the waiting time are minimized while the rate of utilization of the lineacs is maximized. As those goals are contradictory, generating a high quality solution is complex. For example, a palliative
patient should receive a treatment within three days. Considering that the arrival of these patients to the center is unknown the rate of utilization cannot be optimized without exceeding the three days deadline. Furthermore, resource constraints are very tight: the number of patients treated reaches nearly the full capacity. We propose a hybrid algorithm using online and stochastic optimization. Online methods allow scheduling a patient whenever according to the preference of the user and stochastic tools authorize to infer the future in order to take into account prospective patients, especially palliatives whom have very tight deadlines to receive
their treatment. Those techniques provide good results on randomly generated problems as well as real instances. It outperforms current scheduling built by experienced clerks in the center. We are confident about the efficiency of this algorithm since the oncology treatment center can obtain a good inference of future patients. This method will improve the quality of the system both for the patients and for the center.