Abstract
We consider an appointment system where the patients have preferences about the appointment days. A patient may be scheduled on one of the days that is acceptable to her, or be denied appointment. The patient may or may not show up at the appointed time. The net cost is a convex function of the actual number of patients served on a given day. We study the optimal scheduling policy that minimizes the long-run average cost and study its structural properties. We advocate an index policy, which is easy to implement, performs well in comparison with other heuristic policies, and is close to the optimal policy.




Similar content being viewed by others
References
Balasubramanian, H., Biehl, S., Dai, L., Muriel, A.: Dynamic allocation of same-day requests in multi-physician primary care practices in the presence of prescheduled appointments. Health Care Manag. Sci. 17(1), 31–48 (2014)
Blok, H., Spieksma, E.M.: Structures of optimal policies in Markov Decision Processes with unbounded jumps: the state of our art. In: Boucherie, R.J., van Dijk, N.M. (eds.) Markov Decision Processes in Practice, pp. 139–196. Springer, Berlin (2016)
Çayirli, T., Veral, E.: Outpatient scheduling in health care: a review of literature. Prod. Oper. Manag. 12(4), 519–549 (2003)
Dekker, R., Hordijk, A.: Recurrence conditions for average and Blackwell optimality in denumerable state Markov decision chains. Math. Oper. Res. 17(2), 271–289 (1992)
Dekker, R., Hordijk, A., Spieksma, F.M.: On the relation between recurrence and ergodicity properties in denumerable Markov decision chains. Math. Oper. Res. 19(3), 539–559 (1994)
eClinicalWorks: About us. Retrieved from http://www.eclinicalworks.com/about-us/. Accessed 13 Jan 2015
Feldman, J., Liu, N., Topaloglu, H., Ziya, S.: Appointment scheduling under patient preference and no-show behavior. Oper. Res. 62(4), 794–811 (2014)
Green, L.V., Savin, S.: Reducing delays for medical appointments: a queueing approach. Oper. Res. 56(6), 1526–1538 (2008)
Gupta, D., Denton, B.: Appointment scheduling in health care: challenges and opportunities. IIE Trans. 40(9), 800–819 (2008)
Gupta, D., Wang, L.: Revenue management for a primary-care clinic in the presence of patient choice. Oper. Res. 56(3), 576–592 (2008)
Ha, A.Y.: Optimal dynamic scheduling policy for a make-to-stock production system. Oper. Res. 45(1), 42–53 (1997)
Hassin, R., Mendel, S.: Scheduling arrivals to queues: a single-server model with no-shows. Manag. Sci. 54(3), 565–572 (2008)
Kaandorp, G.C., Koole, G.: Optimal outpatient appointment scheduling. Health Care Manag. Sci. 10(3), 217–229 (2007)
Koole, G.: Structural results for the control of queueing systems using event-based dynamic programming. Queueing Syst. 30(3), 323–339 (1998)
Koole, G.: Monotonicity in Markov reward and decision chains: theory and applications. Found. Trends Stoch. Syst. 1, 1–76 (2006)
Kuiper, A., Kemper, B., Mandjes, M.: A computational approach to optimized appointment scheduling. Queueing Syst. 79(1), 5–36 (2015)
LaGanga, L.R., Lawrence, S.R.: Appointment overbooking in health care clinics to improve patient service and clinic performance. Prod. Oper. Manag. 21(5), 874–888 (2012)
Liu, N., Ziya, S., Kulkarni, V.G.: Dynamic scheduling of outpatient appointments under patient no-shows and cancellations. Manuf. Serv. Oper. Manag. 12(2), 347–364 (2010)
Spieksma, F.M.: Geometrically ergodic Markov Chains and the optimal control of queues. PhD thesis, Leiden University (1990) (Available on request from the author)
Talluri, K., van Ryzin, G.: Revenue management under a general discrete choice model of consumer behavior. Manag. Sci. 50(1), 15–33 (2004)
Tijms, H.C.: A First Course in Stochastic Models. Wiley, Chichester (2003)
U.S. Department of Health & Human Services: HITECH Act Enforcement Interim Final Rule. Retrieved from http://www.hhs.gov/ocr/privacy/hipaa/administrative/enforcementrule/hitechenforcementifr.html. Accessed 13 Jan 2015
Wang, W.Y., Gupta, D.: Adaptive appointment systems with patient preferences. Manuf. Serv. Oper. Manag. 13(3), 373–389 (2011)
Zacharias, C., Pinedo, M.: Appointment scheduling with noshows and overbooking. Prod. Oper. Manag. 23(5), 788–801 (2014)
ZocDoc: About us. Retrieved from http://www.zocdoc.com/aboutus. Accessed 13 Jan 2015
Zonderland, M.E., Boucherie, R.J., Hanbali, A.A.: Appointments in care pathways: the \(\text{ Geo }^\text{ x }\)/D/1 queue with slot reservations. Queueing Syst. 79(1), 37–51 (2015)
Acknowledgements
We would like to thank Professor Onno Boxma for his valuable comments on the initial draft. We also want to thank the referees for their careful reading of the paper and extremely useful suggestions that, we believe, have made the paper more readable and more rigorous.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
1.1 Proof of Lemma 3
(a) The result follows since monotonicity is closed under the minimum operator.
(b) We need to show \(T_{12}f(i+1,j)\le T_{12}f(i,j)\) for \(0\le i\le m-1, j\ge 0\). We consider three cases:
Case (b)(1) \(0\le i\le m-2, j\ge 0\). The result follows since monotonicity is closed under the minimum operator.
Case (b)(2) \(i=m-1,0\le j\le m-1\). We need to show \(T_{12}f(m,j)\le T_{12}f(m-1,j)\). We have
Hence
Similarly, we have
Hence
We know \(f(m,j+1)\le f(m,j)\) by \(\textit{Dec}(2) \text { in } S_0\cup S_2\) and \(f(m,j+1)\le f(m-1,j+1)\) by \(\textit{Dec}(1) \text { in } S_0\cup S_1\). Therefore,
Case (b)(3) \(i=m-1,j\ge m\). We need to show \(T_{12}f(m,j)\le T_{12}f(m-1,j)\). We have
Hence
We have
Therefore,
The result follows.
(c) The result follows since monotonicity is closed under the minimum operator.
(d) We need to show \(T_{12}f(i,j+1)\le T_{12}f(i,j)\) for \(i\ge 0,0\le j\le m-1\). We consider three cases:
Case (d)(1) \(i\ge 0, 0\le j\le m-2\). The result follows since monotonicity is closed under the minimum operator.
Case (d)(2) \(0\le i\le m-1,j=m-1\). We need to show \(T_{12}f(i,m)\le T_{12}f(i,m-1)\). We have
Hence
Similarly, we have
Therefore,
We know \(f(i+1,m)\le f(i+1,m-1)\) by \(\textit{Dec}(2) \text { in } S_0\cup S_2\) and \(f(i+1,m)\le f(i,m)\) by \(\textit{Dec}(1) \text { in } S_0\cup S_1\).
Case (d)(3) \(i\ge m,j=m-1\). We need to show \(T_{12}f(i,m)\le T_{12}f(i,m-1)\). We have
Hence
Similarly, we have
Therefore,
The result follows.
(e) We show that \(T_{12}f(i,j)+T_{12}f(i+2,j)\ge 2T_{12}f(i+1,j)\) for \(0\le i\le m-2, 0\le j\le m-1\) and skip the proof that \(T_{12}f(i,j)+T_{12}f(i,j+2)\ge 2T_{12}f(i,j+1)\) for \(0\le i\le m-1, 0\le j\le m-2\) since it is similar. We consider four cases:
Case (e)(1) \(i\le m-3,j\le m-1\). We know
The \(T_{12}\) here coincides with the \(T_{R(I)}\) in Koole [14]. The result follows.
Case (e)(2) \(i\le m-3,j=m\). We show
where we know
The result follows from the \(\textit{Convexity}\) of f.
Case (e)(3) \(i=m-2,j\le m-1\). We show
where we know
We consider two cases:
Case (e)(3)(i) \(f(m-1,j)\ge f(m-2,j+1)\). Then \(T_{12}f(m-2,j)=f(m-2,j+1)\). By the property \(\textit{SuperC}\) of f, we know
Hence \(T_{12}f(m-1,j)=f(m-1,j+1)\). By the \(\textit{Convexity}\) of f, we have
Case (e)(3)(ii) \(f(m-1,j)< f(m-2,j+1)\). Then \(T_{12}f(m-2,j)=f(m-1,j)\). By the \(\textit{Super}\) of f, we have
Case (e)(4) \(i=m-2,j=m\). We show
where we know
The result follows from \(f(m-1,m)\ge f(m,m)\) by \(\textit{Dec}(1) \text { in } S_0\cup S_1\).
(f) We show that \(T_{12}f(i,j)+T_{12}f(i+1,j+1)\ge T_{12}f(i+1,j)+T_{12}f(i,j+1)\) for \(0\le i,j\le m-1\). We consider four cases:
Case (f)(1) \(i\le m-2,j\le m-2\). We know \(T_{12}f(i,j)=\min \{f(i+1,j),f(i,j+1)\}\) coincides with \(T_{R(I)}\) in Koole [14]. The result follows.
Case (f)(2) \(i=m-1,j\le m-2\). We show
where we know
We consider two cases:
Case (f)(2)(i) \(f(m,j)\ge f(m-1,j+1)\). Then \(T_{12}f(m-1,j)=f(m-1,j+1)\). By the \(\textit{Super}\) of f, we know
Case (f)(2)(ii) \(f(m,j)< f(m-1,j+1)\). Then \(T_{12}f(m-1,j)=f(m,j)\). By the \(\textit{SuperC}\) of f, we know
Hence \(T_{12}f(m-1,j+1)=f(m,j+1)\). By the \(\textit{Convexity}\) of f, we have
Case (f)(3) \(i\le m-2,j=m-1\). We show
where we know
We know
Case (f)(4) \(i=m-1,j=m-1\). We show
where we know
We know
(g) We show that \(T_{12}f(i+2,j)+T_{12}f(i,j+1)\ge T_{12}f(i+1,j)+T_{12}f(i+1,j+1)\) for \(0\le i\le m-2, 0\le j\le m-1\). We skip the proof of \(T_{12}f(i+1,j)+T_{12}f(i,j+2)\ge T_{12}f(i,j+1)+T_{12}f(i+1,j+1)\) for \(0\le i\le m-1, 0\le j\le m-2\). We consider four cases:
Case (g)(1) \(i\le m-3,j\le m-2\). We know \(T_{12}f(i,j)=\min \{f(i+1,j),f(i,j+1)\}\) coincides with \(T_{R(I)}\) in Koole [14]. The result follows.
Case (g)(2) \(i=m-2,j\le m-2\). We show
where we know
We consider two cases:
Case (g)(2)(i) \(f(m-1,j+1)<f(m-2,j+2)\). Then \(T_{12}f(m-2,j+1)=f(m-1,j+1)\). We have
Case (g)(2)(ii) \(f(m-1,j+1)\ge f(m-2,j+2)\). Then \(T_{12}f(m-2,j+1)=f(m-2,j+2)\). We have
We have
Case (g)(3) \(i\le m-3,j=m-1\). We show
where we know
We consider two cases:
Case (g)(3)(i) \(f(i+3,m-1)<f(i+2,m)\). Then \(T_{12}f(i+2,m-1)=f(i+3,m-1)\). We have
Hence
Case (g)(3)(ii) \(f(i+3,m-1)\ge f(i+2,m)\). Then \(T_{12}f(i+2,m-1)=f(i+2,m)\). We have
Case (g)(4) \(i=m-2,j=m-1\). We show
where we know
Hence
1.2 Proof of Lemma 4
Due to the symmetry, we show the proof for operator \(T_1\) and skip the proof for operator \(T_2\). Given the properties f has, we know
We prove (f) and skip the rest due to the similarity. We show that \(T_1f(i,j)+T_1f(i+1,j+1)\ge T_1f(i+1,j)+T_1f(i,j+1)\) for any \(0\le i,j\le m-1\). We consider four cases:
Case (f)(1) \(0\le i\le m-2,0\le j\le m-2\). We have
Case (f)(2) \(i=m-1,0\le j\le m-2\). We need to show \(T_1f(m-1,j)+T_1f(m,j+1)\ge T_1f(m,j)+T_1f(m-1,j+1)\). We have
Case (f)(3) \(0\le i\le m-2,j=m-1\). We need to show \(T_1f(i,m-1)+T_1f(i+1,m)\ge T_1f(i+1,m-1)+T_1f(i,m)\). We have
Case (f)(4) \(i=m-1,j=m-1\). We need to show \(T_1f(m-1,m-1)+T_1f(m,m)\ge T_1f(m,m-1)+T_1f(m-1,m)\). This follows because
1.3 Proof of Lemma 6
We know \(T_cf(i,j)=c(i)+f(j,0)\) and c(x) is a convex function which achieves its minimum at \(x=m\), hence (a) and (b) follow. For (c), if f is \(\textit{Inc}(1) \text { in } S_2\cup S_3\), then \(f(i+1,j)\ge f(i,j), \forall i\ge m, j\ge 0\). For any \(j\ge m\), we have \(T_cf(i,j+1)=c(i)+f(j+1,0)\ge c(i)+f(j,0)=T_cf(i,j)\). Hence \(T_cf\) is \(\textit{Inc}(2) \text { in } S_1\cup S_3\). For (d), if f is \(\textit{Dec}(1) \text { in } S_0\cup S_1\), then \(f(i+1,j)\le f(i,j), \forall i\le m-1, j\ge 0\). For any \(j\le m-1\), we have \(T_cf(i,j+1)=c(i)+f(j+1,0)\le c(i)+f(j,0)=T_cf(i,j)\). Hence \(T_cf\) is \(\textit{Dec}(2) \text { in } S_0\cup S_2\). For (e), the Convexity of \(T_cf\) follows directly from the Convexity of f. (f) is trivially shown given the definition of the operator \(T_c\). For (g), we have
and
Rights and permissions
About this article
Cite this article
Zhang, Y., Kulkarni, V.G. Two-day appointment scheduling with patient preferences and geometric arrivals. Queueing Syst 85, 173–209 (2017). https://doi.org/10.1007/s11134-016-9506-x
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11134-016-9506-x