Skip to main content

Demand modelling for emergency medical service system with multiple casualties cases: k-inflated mixture regression model


In most of the literature on emergency medical service (EMS) system design and analysis, arrivals of EMS calls are assumed to follow Poisson process. However, it is not uncommon for real-world EMS systems to experience batch arrivals of EMS requests, where a single call involves more than one patient. Properly capturing such batch arrivals is needed to enhance the quality of analyses, thereby improving the fidelity of a resulting system design. This paper proposes a spatio-temporal demand model that incorporates batch arrivals of EMS calls. Specifically, we construct a spatio-temporal compound Poisson process which consists of a call arrival model and call size model. We build our call arrival model by combining two models available in the existing EMS demand modeling literature—artificial neural network and spatio-temporal Gaussian mixture model. For the call size model, we develop a k-inflated mixture regression model. This model reflects the characteristics of EMS call arrivals that most calls involve one patient while some calls involve multiple patients. The utility of the proposed EMS demand model is illustrated by a probabilistic ambulance location model, where we show ignoring batch arrivals leads to overestimation of ambulance availability.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12


  1. 1.

    We process \(\varvec{X}_t\) to a suitable form for each models. For example, we use artificial neural network (ANN) to model \(\varLambda _t\) and \(\varvec{X}_t\) is processed to three discrete variables; month-of-year, day-of-week, hour-of-day.

  2. 2.

    We used the EMS call data from the city of Daejeon, Korea, for the period of January 1st, 2011 to December 31st, 2012, and test each model for their forecasting accuracy(Lee and Lee 2016). ANN showed better performance than the other two methods in forecasting length of 1 hour, 1 day, 3 days, 1 week. Results from these experiments can be supplied upon request.

  3. 3.

    \(\varvec{\omega }_t = \left( \omega _{1}(\varvec{X}_t), ..., \omega _{C}(\varvec{X}_t)\right)\).

  4. 4.

    Equations for implementing an EM algorithm for geometric and negative binomial distribution are provided in “Appendix A”.

  5. 5.

    For the GA algorithm, we set the population size at 2000, elite population 5%, crossover probability to be 60%, and mutation probability to be 5%.

  6. 6.

    \(e = |\)loss probability estimated from hypercube queuing - loss probability (fraction of lost calls) measured in simulation|.


  1. Alanis R, Ingolfsson A, Kolfal B (2013) A Markov chain model for an EMS system with repositioning. Prod Oper Manag 22(1):216–231

    Article  Google Scholar 

  2. Azeez Jazeem, Aravindhar D John (2015) Hybrid approach to crime prediction using deep learning. In: 2015 International conference on advances in computing, communications and informatics (ICACCI), pp. 1701–1710. IEEE

  3. Baker JR, Fitzpatrick KE (1986) Determination of an optimal forecast model for ambulance demand using goal programming. J Oper Res Soc 37(11):1047–1059

    Article  Google Scholar 

  4. Bélanger V, Ruiz A, Soriano P (2019) Recent optimization models and trends in location, relocation, and dispatching of emergency medical vehicles. Eur J Oper Res 272(1):1–23

    MathSciNet  Article  Google Scholar 

  5. Bishop CM (2006) Pattern recognition and machine learning. Springer, Berlin

    MATH  Google Scholar 

  6. Brotcorne L, Laporte G, Semet F (2003) Ambulance location and relocation models. Eur J Oper Res 147(3):451–463

    MathSciNet  Article  Google Scholar 

  7. Budge S, Ingolfsson A, Erkut E (2009) Approximating vehicle dispatch probabilities for emergency service systems with location-specific service times and multiple units per location. Oper Res 57(1):251–255

    Article  Google Scholar 

  8. Channouf N, LEcuyer P, Ingolfsson A, Avramidis AN (2007) The application of forecasting techniques to modeling emergency medical system calls in Calgary, Alberta. Health Care Manag Sci 10(1):25–45

    Article  Google Scholar 

  9. Chen AY, Lu T-Y, Ma MH-M, Sun W-Z (2016) Demand forecast using data analytics for the pre-allocation of ambulances. IEEE J Biomed Health Inf 20(4):1178–1187

    Article  Google Scholar 

  10. Finkelman MD, Green JG, Gruber MJ, Zaslavsky AM (2011) A zero-and k-inflated mixture model for health questionnaire data. Stat Med 30(9):1028–1043

    MathSciNet  Article  Google Scholar 

  11. Freitas AV, Afreixo V, Cruz SE (2013) Mixture models of geometric distributions in genomic analysis of inter-nucleotide distances. Stat Optim Inf Comput 1(1):8–28

    Article  Google Scholar 

  12. Fritsch Stefan, Guenther Frauke (2016) Package neuralnet. The Comprehensive R Archive Network

  13. Fujiwara O, Makjamroen T, Gupta KK (1987) Ambulance deployment analysis: a case study of Bangkok. Eur J Oper Res 31(1):9–18

    Article  Google Scholar 

  14. Furman E (2007) On the convolution of the negative binomial random variables. Stat Probabl Lett 77(2):169–172

    MathSciNet  Article  Google Scholar 

  15. Gendreau M, Laporte G, Semet F (2006) The maximal expected coverage relocation problem for emergency vehicles. J Oper Res Soc 57(1):22–28

    Article  Google Scholar 

  16. Goldberg J, Dietrich R, Chen JM, Mitwasi MG, Valenzuela T, Criss E (1990) Validating and applying a model for locating emergency medical vehicles in Tuczon, AZ. Eur J Oper Res 49(3):308–324

    Article  Google Scholar 

  17. Goldberg JB (2004) Operations research models for the deployment of emergency services vehicles. EMS Manag J 1(1):20–39

    Google Scholar 

  18. Jarvis JP (1985) Approximating the equilibrium behavior of multi-server loss systems. Manage Sci 31(2):235–239

    Article  Google Scholar 

  19. Jin LH, Taesik L (2016) Comparison of spatio-temporal emergency medical demand forecasting model for emergency medical system operation planning. In: Proceedings of the Korea Institute of Industrial Engineers fall conference, pp. 1864–1882

  20. Kamenetzky RD, Shuman LJ, Wolfe H (1982) Estimating need and demand for prehospital care. Oper Res 30(6):1148–1167

    Article  Google Scholar 

  21. Kang H-W, Kang H-B (2017) Prediction of crime occurrence from multi-modal data using deep learning. PLoS ONE 12(4):e0176244

    Article  Google Scholar 

  22. Larson RC (1974) A hypercube queuing model for facility location and redistricting in urban emergency services. Comput Oper Res 1(1):67–95

    Article  Google Scholar 

  23. Larson RC (1975) Approximating the performance of urban emergency service systems. Oper Res 23(5):845–868

    Article  Google Scholar 

  24. Matteson DS, McLean MW, Woodard DB, Henderson SG et al (2011) Forecasting emergency medical service call arrival rates. Ann Appl Stati 5(2B):1379–1406

    MathSciNet  Article  Google Scholar 

  25. McCarthy ML, Zeger SL, Ding R, Aronsky D, Hoot NR, Kelen GD (2008) The challenge of predicting demand for emergency department services. Acad Emerg Med 15(4):337–346

    Article  Google Scholar 

  26. Najafabadi ATP, MohammadPour S (2017) A k-inflated negative binomial mixture regression model: application to rate–making systems. Asia-Pacific J Risk Insur 12(2).

    Article  Google Scholar 

  27. Nakaya T, Yano K (2010) Visualising crime clusters in a space-time cube: an exploratory data-analysis approach using space-time kernel density estimation and scan statistics. Trans GIS 14(3):223–239

    Article  Google Scholar 

  28. Naoum-Sawaya J, Elhedhli S (2013) A stochastic optimization model for real-time ambulance redeployment. Comput Oper Res 40(8):1972–1978

    Article  Google Scholar 

  29. Papier F, Thonemann UW (2008) Queuing models for sizing and structuring rental fleets. Transp Sci 42(3):302–317

    Article  Google Scholar 

  30. Peleg K, Pliskin JS (2004) A geographic information system simulation model of ems: reducing ambulance response time. Am J Emerg Med 22(3):164–170

    Article  Google Scholar 

  31. Puig P, Valero J (2006) Count data distributions: some characterizations with applications. J Am Stat Assoc 101(473):332–340

    MathSciNet  Article  Google Scholar 

  32. Repede JF, Bernardo JJ (1994) Developing and validating a decision support system for locating emergency medical vehicles in Louisville, Kentucky. Eur J Oper Res 75(3):567–581

    Article  Google Scholar 

  33. Sarabia JM, Prieto F, Trueba C (2012) The n-fold convolution of a finite mixture of densities. Appl Math Comput 218(19):9992–9996

    MathSciNet  MATH  Google Scholar 

  34. Schmid V (2012) Solving the dynamic ambulance relocation and dispatching problem using approximate dynamic programming. Eur J Oper Res 219(3):611–621

    MathSciNet  Article  Google Scholar 

  35. Setzler H, Saydam C, Park S (2009) Ems call volume predictions: a comparative study. Comput Oper Res 36(6):1843–1851

    Article  Google Scholar 

  36. Sung I, Lee T (2018) Scenario-based approach for the ambulance location problem with stochastic call arrivals under a dispatching policy. Flex Serv Manuf J 30(1–2):153–170

    Article  Google Scholar 

  37. Taddy MA, Kottas A et al (2012) Mixture modeling for marked Poisson processes. Bayesian Anal 7(2):335–362

    MathSciNet  Article  Google Scholar 

  38. Vile JL, Gillard JW, Harper PR, Knight VA (2012) Predicting ambulance demand using singular spectrum analysis. J Oper Res Soc 63(11):1556–1565

    Article  Google Scholar 

  39. Wei LSS, Sian NY, Rajagopal LM, Yng NY, Hock OME (2016) Ambulance deployment under demand uncertainty. J Adv Manag Sci 4(3):187–194

    Google Scholar 

  40. Winters PR (1960) Forecasting sales by exponentially weighted moving averages. Manage Sci 6(3):324–342

    MathSciNet  Article  Google Scholar 

  41. Wolfe P (1971) Convergence conditions for ascent methods. ii: Some corrections. SIAM review 13(2):185–188

    MathSciNet  Article  Google Scholar 

  42. Zhou Zhengyi, Matteson David S (2016) Temporal and spatio-temporal models for ambulance demand. Healthcare Analytics: From Data to Knowledge to Healthcare Improvement, pp. 389–412

  43. Zhou Z, Matteson DS, Woodard DB, Henderson SG, Micheas AC (2015) A spatio-temporal point process model for ambulance demand. J Am Stat Assoc 110(509):6–15

    MathSciNet  Article  Google Scholar 

Download references


This research was supported by the “Decision Models in Behavioral Operations Research” through the National Research Foundation of Korea(NRF), funded by the Ministry of Science and ICT (2019-R1A2C108830212).

Author information




All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by HL and TL. The first draft of the manuscript was written by HL and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Taesik Lee.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


Appendix A: First and second order derivatives for negative binomial distribution and geometric distribution

Let us denote the two parameters of a negative binomial distribution by \(\xi\) (number of failures) and \(\delta\) (probability of success in a Bernoulli trial). Then, the first and second order derivatives of \(Q_1\) and \(Q_2\) for the negative binomial distribution are

$$\begin{aligned} \frac{\partial Q_1}{\partial \alpha _{ml}}= & {} \sum _{J}V_{i,t,l}[z_{i,t,m}-\pi _{i,t,m}], \\ \frac{\partial Q_2}{\partial \xi _m}= & {} \sum _{J}z_{i,t,m}[\varGamma (b_{i,t}+\xi _m-1)-\varGamma (\xi _m)+log(1-\delta _m)], \\ \frac{\partial Q_2}{\partial \delta _m}= & {} \sum _{J}z_{i,t,m}[\frac{b_{i,t}-1}{\delta _m}-\frac{\xi _m}{1-\delta _m}], \\ {\frac{\partial ^2 Q_1}{\partial \alpha _{ml}^2}}= & {} \sum _{J}V_{i,t,l}^2\pi _{im}, \\ {\frac{\partial ^2 Q_2}{\partial \xi _m^2}}= & {} \sum _{J}z_{i,t,m}[\psi (b_{i,t}+\xi _m-1)-\psi (\xi _j)], \\ {\frac{\partial ^2 Q_2}{\partial \delta _m^2}}= & {} \sum _{J}z_{i,t,m}\left[ \frac{1-b_{i,t}}{\delta _m^2}-\frac{\xi _m}{(1-\delta _m)^2}\right] , \end{aligned}$$

where \(\alpha _{ml}\) indicate the coefficient for \(l^{th}\) variables in \(V_{i,t}\), \(\varGamma\) denotes gamma function, and \(\psi\) represent digamma function \(\psi (x) = \frac{d}{dx}ln(\varGamma (x))\).

For geometric distribution, we use \(\zeta\) to denote its parameter – success probability in Bernoulli trials. Its first and second derivatives are

$$\begin{aligned} \frac{\partial Q_1}{\partial \alpha _{ml}}= & {} \sum _{J}V_{i,t,l}[z_{i,t,m}-\pi _{i,t,m}], \\ \frac{\partial Q_2}{\partial \zeta _m}= & {} \sum _{J}z_{i,t,m}[\frac{1}{(1-\zeta _m)\zeta _m}-\frac{b_{i,t}}{1-\zeta _m}], \\ {\frac{\partial ^2 Q_1}{\partial \alpha _{ml}^2}}= & {} \sum _{J}V_{i,t,l}^2\pi _{i,t,m}, \\ {\frac{\partial ^2 Q_2}{\partial \zeta _m^2}}= & {} \sum _{J}z_{i,t,m}\left[ \frac{2\zeta _m-1}{(1-\zeta _m)^2\zeta _m^2}-\frac{b_{i,t}}{(1-\zeta _m)^2}\right] . \end{aligned}$$

Appendix B: Batch arrival hypercube queuing approximation

In this section, we briefly introduce the hypercube queuing model (Larson 1975; Jarvis 1985; Budge et al. 2009), and explain how we modify the model to incorporate batch arrivals. Consider a network of N servers that serve D demand nodes. For each demand node, servers are ordered by their preference rank by the node. Let \(g_{ji}\) denote a dispatching probability of server j to demand node i when i is a \(k^{th}\) preferred server. Then, \(g_{ji}\) is approximated as,

$$\begin{aligned} g_{ji} \approx Q(N,\rho ,k)(1-\rho _j)\prod _{l=1}^{k-1}\rho _{a_{il}}. \end{aligned}$$

In Equation (B.1), \(\rho\) represents the system-wise busy fraction, \(\rho _j\) denotes the busy fraction of server j, and \(a_{il}\) denotes the index of \(l^{th}\) preferred server for demand node i. The system-wise busy fraction \(\rho\) is determined as \(\lambda \tau / N\), where \(\lambda\) is the system-wise arrival rate and \(\tau\) is the average service time. The essence of this approximation is the correction factor \(Q(N,\rho ,k)\). \(Q(N,\rho ,k)\) is introduced in Equation (B.1) in order to compensate for the errors due to the dependence of servers in the queuing network. Larson (1975) proposed \(Q(N,\rho ,k)\) for the hypercube queue as shown below:

$$\begin{aligned} Q(N,\rho ,k) = \frac{(N-k)!P_0}{N!(1-\rho (1-P_N))(1-P_N)^{k-1}}\sum _{u=k-1}^{N-1}\frac{(N-u)(N^u\rho ^{u-k+1})}{(u-k+1)!}, \end{aligned}$$

where \(P_0\) and \(P_N\) denote the idle probability and loss probability of an M/M/N/N queue. Since the probability that u servers are busy, \(P_u\), in an M/M/N/N queue is \(P_u = \frac{(N\rho )^u}{u!}*P_0\), we can rewrite Equation B.2 as

$$\begin{aligned} Q(N,\rho ,k) = \frac{(N-k)!}{N!(1-\rho (1-P_N))(\rho (1-P_N))^{k-1}}\sum _{u=k-1}^{N-1}\frac{(N-u)u!}{(u-k+1)!}P_u. \end{aligned}$$

To account for batch arrivals, we use \(P_u\) from \(M^X/G/N/N\) queue (Papier and Thonemann 2008), instead of from M/M/N/N queue:

$$\begin{aligned} P_u&= \frac{e^{N\rho }\sum _{a=0}^u\frac{(N\rho )^{a}}{a!}\varPhi ^{(a)}(u)}{\sum _{u=0}^N e^{N\rho }\sum _{a=0}^u\frac{(N\rho )^{a}}{a!}\varPhi ^{(a)}(u)} \nonumber \\&= P_0 * \sum _{a=0}^u\frac{(N\rho )^{a}}{a!}\varPhi ^{(a)}(u) \end{aligned}$$

where \(\varPhi ^{(a)}\) denotes \(a^{\text {th}}\) convolution of batch size distribution \(\varPhi\). That is, \(\varPhi ^{(a)}(u)\) represents the probability that the sum of a batches sampled from distribution \(\varPhi\) is u. Then, a generalized correction factor \(Q^*(N,\rho ,k)\) is

$$\begin{aligned} Q^*(N,\rho ,k)&= \frac{(N-k)!P_0}{N!(1-\rho (1-P_N))(1-P_N)^{k-1}} \nonumber \\&\qquad \sum _{u=k-1}^{N-1}\frac{(N-u)u!}{(u-k+1)!}\sum _{a=0}^{u}\frac{N^a\rho ^{a-k+1}}{a!}\varPhi ^{(a)}(u), \end{aligned}$$

It is easy to show that the generalized correction factor reduces to the single arrival correction factor, Equation (B.2).

In Equation (B.5), the most challenging part is the calculation of \(\varPhi ^{(a)}(u)\). Note that we use a mixture model for our call size model, i.e., \(\varPhi (u)=\sum _{m\in {1,\ldots ,M}}\pi _m\phi _m(u)\), and based on the results from Sarabia et al. (2012), we derive \(n^{\text {th}}\) convolution of the mixture distribution as

$$\begin{aligned} \varPhi ^{(n)}(u)=\sum _{(n_1,\ldots ,n_M)\in C_n}\frac{n!}{n_1!,\ldots ,n_M!}\pi _1^{n_1}...\pi _M^{n_M}\phi _1^{(n_1)} \otimes ...\otimes \phi _M^{(n_M)}(u), \end{aligned}$$

where \(C_n\) is a set of combinations for n batches’ mixture membership.

In our model, we use a geometric distribution for \(\phi _m\). Since the sum of random variables from geometric distribution follows a negative binomial distribution, Equation (B.6) becomes,

$$\begin{aligned} \varPhi ^{(n)}(u)&= \nonumber \\&\quad \sum _{(n_1,\ldots ,n_M)\in C_n}\frac{n!}{n_1!,\ldots ,n_M!}\pi _1^{n_1}...\pi _M^{n_M}NB(n_1,1-\zeta _1) \otimes ...\otimes NB(n_M,1-\zeta _M)(u). \end{aligned}$$

Furman (2007) provides an approximation to evaluate the convolution of a negative binomial distribution. (refer to Theorem 2. of Furman (2007)).

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Lee, H., Lee, T. Demand modelling for emergency medical service system with multiple casualties cases: k-inflated mixture regression model. Flex Serv Manuf J (2021).

Download citation


  • Demand modeling
  • Batch arrival
  • Spatio-temporal variability
  • Emergency medical service system
  • Mixture model