A New Appointment Rule for a Single-Server, Multiple-Customer Service System Kum Khiong Yang, Mun Ling Lau, Ser Aik Quek Department of Decision Sciences, National University of Singapore, Kent Ridge, Singapore 119260 Received September 1996; revised April 1997; accepted 31 October 1997 Abstract: This paper proposes a new appointment rule for the single-server, multiple- customer service system. Unlike previous appointment rules, which perform well only in specific service environments, the new rule can be parameterized to perform well in different service environments. The new appointment rule is presented as a mathematical function of four environmental parameters, namely, the coefficient of variation of the service time, the percentage of customers’ no-shows, the number of appointments per service session, and the cost ratio between the server’s idle and customers’ waiting cost per unit time. Once the values of these environmental parameters are estimated, the new appointment rule can be parameterized to perform well. The results show that new rule performs either as well as or better than existing appointment rules in a wide range of service environments.  1998 John Wiley & Sons, Inc. Naval Research Logistics 45: 313–326, 1998 Keywords: scheduling; appointment system; simulation 1. INTRODUCTION There are many examples of a single-server, multiple-customer service system. These examples include medical clinics, dental offices, law firms, and hairdressing saloons where each server independently services a group of loyal customers. However, in spite of the diversity in the types of services offered, many of these service systems share a common goal to improve the value and services offered to their customers. One approach in this respect is the use of a good appointment rule to manage the total server’s idle and customers’ waiting costs. Many past studies [1–3, 8, 18] have proposed appointment rules which perform well in specific service environments, but poorly in other service environments. To address the limitation of these past studies, Ho and Lau [ 9 ] proposed a procedure for choosing the ‘‘best’’ appointment rule for different service environments from a set of eight ‘‘best’’ rules. The use of Ho and Lau’s procedure [9] is, however, constrained to the service environments examined in their study. For an environment that is different from those tested in their study, Ho and Lau [9] suggested additional computer simulation to find the appropriate appointment rule. Many of their proposed appointment rules also perform well in a narrow range of cost Correspondence to: Kum Khiong Yang  1998 by John Wiley & Sons, Inc. CCC 0894-069X/98/030313-14 972 / 8m21$$0972 01-30-98 10:10:14 nra W: Nav Res