ORAHS2024
Abstract Submission

292. Appointment Scheduling with Updates: An Exact and Optimal Approach

Contributed abstract in session MC-5: (Multi)Appointment /1, stream Regular talks.

Monday, 11:00-12:30
Room: Room S6

Authors (first author is the speaker)

1. Alex Kuiper
Business Analytics, University of Amsterdam
2. Roshan Mahes
University of Amsterdam

Abstract

Traditionally, appointment schedules have been determined by minimizing a specific cost function consisting of clients’ waiting times and server idling. Under stochastic service times, this intra-day problem is predominantly studied in a static sense, assuming that a schedule is fixed. However, technological advancements have opened up the possibility of communicating with clients during the day. Using such channels allows for updating appointment schedules on the fly, for example, by postponing appointments in case of overcrowding. Yet, while its promise is clear, sending updates at the wrong moments, or too many, may be ineffective or unrealistic, potentially causing client confusion and frustration.

Three rescheduling paradigms are introduced that can be solved to optimality via dynamic programming. The experiments indicate that total costs are significantly decreasing, even with relatively few updates. The benefits concentrate on increasing utilization (less idling). The magnitude of the decrease is moderated by the length of the grace period, which accounts for not sending updates to clients possibly underway. If the rescheduling decisions are chosen a priori, then they should be equally spread throughout the session to be most effective. The best performance is, however, obtained under the third paradigm of dynamic rescheduling in the case the desired maximum of clients in the system is exceeded.

Keywords

Status: accepted

Back to the list of papers