Scheduling and Sequencing Arrivals to a Stochastic Service System
Abstract
Optimization of scheduled arrival times to an appointment system is approached from the perspectives of both queueing and scheduling theory. The appointment system is modeled as a one-server, first-come-first-served, transient queue with independent, distinctly distributed service times and no-show rates. If a customer does show, it is assumed to be punctual. The cost of operating the appointment system is a convex combination of customers' waiting times and the server's overtime. While techniques for finding the optimal static and dynamic schedules of arrivals have been proposed by other researchers, they mainly have focused on identical customers and strictly punctual arrivals. This effort provides substantially more efficient solution methods, addresses a more general cost function, allows for no-shows and non-identical service distributions, and applies either when arrivals are constrained to lattice points or when they are unconstrained. Because customers are not indistinguishable, this effort also provides heuristics for determining optimal customer order. The effort concentrates on medical scheduling examples but is applicable to any appointment scheduling operation. Further, the proposed techniques apply to any convex, submodular function.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1997
- Accession Number
- ADA332136
Entities
People
- Peter M. Bosch
Organizations
- Air Force Institute of Technology