Simulation and Optimization in Hospital Planning and Scheduling

基於仿真與優化的醫院資源規劃和調度研究

Student thesis: Doctoral Thesis

View graph of relations

Author(s)

Detail(s)

Awarding Institution
Supervisors/Advisors
Award date17 Oct 2017

Abstract

Over the past years a vast amount of literature has emerged on the problem of health care planning and control, to gain efficiency and effectiveness in the outpatient environment. This is warranted by an increase in demand for health care and increasing expenditures, while a substantial part of the expenditures is made in hospitals. Planning and control in hospitals is fragmented, because large care providers such as hospitals generally consist of autonomously managed departments. This is reflected in research, with many research efforts focused on a single care path component. Examples include outpatient appointment scheduling (AS) or operating room scheduling (OR). Patient scheduling and/or resource allocation is often optimized for one of these components, but not for the entire patient care path. Complexity is often one of the key impediments to integration. The OR is often considered a key resource since it involves high costs and is thus an interesting unit to optimize. Although some studies include downstream bed occupancy in their planning and scheduling models, most of them use a scope that excludes previous or subsequent episodes of the patient's entire care path. Through a series of three related research efforts, this dissertation investigates how the key resource - the specialist, or surgeon - should be allocated to his/her activities, in order to accommodate a chain of appointments in the patients' heterogeneous care paths. Variability in resource availability plays a critical role in each of these efforts.

In the first research effort an algorithm is developed to allocate specialists to their day-to-day activities, in a non-cyclic way. With a number activities to prioritize, the aim is to allocate those activities that have the most detrimental effect in terms of variability, were they not allocated in the upcoming roster. The second research effort builds on this, but takes a more stringent approach of optimization through mathematical programming, with additional real-life constraints. The third research effort is an investigation of simultaneous allocation of surgeons and patients to OR-blocks. As with the previous two research efforts, there is a focus on keeping variability low, in this case by balancing it among all OR-blocks of a single day. In the second and third research effort mathematical programming is the main methodology used, while discrete-event simulation is used in all three research efforts to assess the performance of each of the developed solutions.

The results of this research are supportive of a care path approach to key resource allocation. Moreover, traditional cyclic rosters have their advantages, but are rigid without mechanisms for dealing with variability in the availability of key resources. Dynamic resource allocation methods that actively manage supply-side variability enable timely patient appointments throughout their care paths.

    Research areas

  • Planning, Scheduling, Hospital, Optimization, Discrete-event simulation