Abstract
Queueing systems experienced in real-life situations are very often influenced by negative arrivals which are independent of service process and cause the elimination of jobs from the system. Such a scenario occurs in computer network and telecommunication systems where an attack by a malicious virus results in the removal of some or all data files from the system. Along this direction many authors have proposed various killing processes in the past. This paper unifies different killing mechanisms into the classical single server queue having infinite capacity, where arrival occurs as renewal process with exponential service time distribution. The system is assumed to be affected by negative customers as well as disasters. The model is investigated in steady-state in a very simple and elegant way by means of supplementary variable and difference equation technique. The distribution of system-content for the positive customers is derived in an explicit form at pre-arrival and random epochs. The influence of different parameters on the system performance are also examined.
Similar content being viewed by others
References
Abolnikov, L., Dukhovny, A.: Markov chains with transition delta-matrix: ergodicity conditions, invariant probability measures and applications. Int. J. Stoch. Anal. 4(4), 333–355 (1990)
Artalejo, J.R.: G-networks: a versatile approach for work removal in queueing networks. Eur. J. Oper. Res. 126(2), 233–249 (2000)
Atencia, I., Moreno, P.: The discrete-time \(Geo/Geo/1\) queue with negative customers and disasters. Comput. Oper. Res. 31(9), 1537–1548 (2004)
Barbhuiya, F., Gupta, U.: A difference equation approach for analysing a batch service queue with the batch renewal arrival process. J. Differ. Equ. Appl. 25(2), 233–242 (2019)
Barbhuiya, F., Gupta, U.: Discrete-time queue with batch renewal input and random serving capacity rule: \({GI^{X}/Geo^{Y}/1}\). Queueing Syst. 91(3–4), 347–365 (2019)
Barbhuiya, F., Kumar, N., Gupta, U.: Batch renewal arrival process subject to geometric catastrophes. Methodol. Comput. Appl. Probab. 21(1), 69–83 (2019)
Chae, K.C., Park, H.M., Yang, W.S.: A \(GI/Geo/1\) queue with negative and positive customers. Appl. Math. Model. 34(6), 1662–1671 (2010)
Chao, X., Miyazawa, M., Pinedo, M.: Queueing Networks: Customers, Signals and Product Form Solutions. Wiley, New York (1999)
Chen, A., Renshaw, E.: The \(M/M/1\) queue with mass exodus and mass arrivals when empty. J. Appl. Probab. 34(1), 192–207 (1997)
Dharmaraja, S., Kumar, R.: Transient solution of a markovian queuing model with heterogeneous servers and catastrophes. Opsearch 52(4), 810–826 (2015)
Do, T.V.: Bibliography on G-networks, negative customers and applications. Math. Comput. Model. Int. J. 53(1–2), 205–212 (2011)
Do, T.V.: An initiative for a classified bibliography on G-networks. Perform. Eval. 68(4), 385–394 (2011)
Gelenbe, E.: Random neural networks with negative and positive signals and product form solution. Neural Comput. 1(4), 502–510 (1989)
Harrison, P.G., Pitel, E.: Sojourn times in single-server queues by negative customers. J. Appl. Probab. 30(4), 943–963 (1993)
Harrison, P.G., Pitel, E.: The \({M/G/1}\) queue with negative customers. Adv. Appl. Probab. 28(2), 540–566 (1996)
Jain, G., Sigman, K.: A Pollaczek-Khintchine formula for \({M/G/1}\) queues with disasters. J. Appl. Probab. 33(4), 1191–1200 (1996)
Jolai, F., Asadzadeh, S., Taghizadeh, M.: Performance estimation of an email contact center by a finite source discrete time \({Geo/Geo/1}\) queue with disasters. Comput. Ind. Eng. 55(3), 543–556 (2008)
Kalidass, K., Gopinath, S., Gnanaraj, J., Ramanath, K.: Time dependent analysis of an \({M/M/1/N}\) queue with catastrophes and a repairable server. Opsearch 49(1), 39–61 (2012)
Kim, B.K., Lee, D.H.: The \({M/G/1}\) queue with disasters and working breakdowns. Appl. Math. Model. 38(5–6), 1788–1798 (2014)
Koh, S., Chin, C., Tan, Y., Teoh, L., Pooi, A., Goh, Y.: Stationary queue length of a single-server queue with negative arrivals and nonexponential service time distributions. In: MATEC Web of Conferences, vol. 189, p. 02006. EDP Sciences (2018)
Laxmi, P.V., Soujanya, M.: Perishable inventory system with service interruptions, retrial demands and negative customers. Appl. Math. Comput. 262, 102–110 (2015)
Laxmi, P.V., Soujanya, M.: Retrial inventory system with negative customers and service interruptions. Opsearch 52(2), 212–224 (2015)
Lee, D.H., Park, H.M., Lim, D.E.: Queue length and busy period analysis for the \({M/G/1}\) queue with negative arrivals. Indian J. Sci. Technol. 8, 34 (2015)
Park, H.M., Yang, W.S., Chae, K.C.: Analysis of the \({GI/Geo/1}\) queue with disasters. Stoch. Anal. Appl. 28(1), 44–53 (2009)
Park, H.M., Yang, W.S., Chae, K.C.: The \({Geo/G/1}\) queue with negative customers and disasters. Stoch. Models 25(4), 673–688 (2009)
Rajadurai, P., Chandrasekaran, V., Saravanarajan, M.: Analysis of an \({M^{[X]}/G/1}\) unreliable retrial G-queue with orbital search and feedback under Bernoulli vacation schedule. Opsearch 53(1), 197–223 (2016)
Rakhee, Sharma, G., Priya, K.: Analysis of G-queue with unreliable server. Opsearch 50(3), 334–345 (2013)
Singh, G., Gupta, U., Chaudhry, M.L.: Detailed computational analysis of queueing-time distributions of the \({BMAP/G/1}\) queue using roots. J. Appl. Probab. 53(4), 1078–1097 (2016)
Towsley, D., Tripathi, S.K.: A single server priority queue with server failures and queue flushing. Oper. Res. Lett. 10(6), 353–362 (1991)
Wang, J., Huang, Y., Dai, Z.: A discrete-time on–off source queueing system with negative customers. Comput. Ind. Eng. 61(4), 1226–1232 (2011)
Wang, J., Zhang, P.: A single-server discrete-time retrial G-queue with server breakdowns and repairs. Acta Math. Appl. Sin. Engl. Ser. 25(4), 675–684 (2009)
Yang, S., Wu, J., Liu, Z.: An \({M^{[X]}/G/1}\) retrial G-queue with single vacation subject to the server breakdown and repair. Acta Math. Appl. Sin. Engl. Ser. 29(3), 579–596 (2013)
Yang, W.S., Chae, K.C.: A note on the \({GI/M/1}\) queue with poisson negative arrivals. J. Appl. Probab. 38(4), 1081–1085 (2001)
Yi, X.W., Kim, J.D., Choi, D.W., Chae, K.C.: The \({Geo/G/1}\) queue with disasters and multiple working vacations. Stoch. Models 23(4), 537–549 (2007)
Zhang, M., Liu, Q.: An \({M/G/1}\) G-queue with server breakdown, working vacations and vacation interruption. Opsearch 52(2), 256–270 (2015)
Zhou, W.H.: Performance analysis of discrete-time queue \({GI/G/1}\) with negative arrivals. Appl. Math. Comput. 170(2), 1349–1355 (2005)
Acknowledgements
Nitin Kumar and F. P. Barbhuiya are grateful to Indian Institute of Technology Kharagpur, India for the financial support. The authors are thankful to the editor and two anonymous referees for their valuable remarks and suggestions which led to the paper in current form.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
Corollary 1
The equation\(A^*(\mu - \mu z)p+ A^*(\mu - \mu z)q_1z^2-z=0\)has only one root in (0, 1) iff\(\frac{\lambda (p-q_1)}{\mu }<1\).
Proof
Let \(f(z)=-z\) and \(g(z)= A^*(\mu - \mu z)p+ A^*(\mu - \mu z)q_1z^2\). Let \(H(z)=A^*(\mu - \mu z)\), it is analytic inside the unit circle and can be written as \(H(z)=\sum _{i=0}^{\infty }h_iz^i\). The coefficients are such that \(h_i \ge 0~~ \forall ~ i \ge 0\). Consider the circle \(|z|=1- \epsilon \) where \(\epsilon >0\) is a very small quantity. Now, \(|f(z)|= 1- \epsilon \) and
From (18) it can be observed that \(|g(z)|< |f(z)|\) iff \((\frac{\mu }{\lambda }+2q_1)>1\) i.e., \(\frac{\lambda (p-q_1)}{\mu }<1\). Thus using Rouch\(\acute{e}\)’s theorem, f(z) and \(f(z)+g(z)\) have equal number of zeroes inside \(|z|=1\) and hence \(f(z)+g(z)\) has only one root in the interval (0, 1) under the condition \(\frac{\lambda (p-q_1)}{\mu }<1\). \(\square \)
Corollary 2
The equation\(A^*(\mu - \mu z)p-z=0\)has only one root in (0, 1) iff\(\frac{\mu p}{\lambda }>1\).
Proof
As before let \(f(z)=-z\) and \(g(z)= A^*(\mu - \mu z)p\) and consider \(|z|=1- \epsilon \). Since \(|f(z)|= 1- \epsilon \) and \(|g(z)|= p- \frac{\epsilon p \mu }{\lambda }< 1-\frac{\epsilon p \mu }{\lambda }\), this suggests that \(f(z)+g(z)\) has only one root in (0, 1) under the condition \(\frac{\mu p}{\lambda }>1\). \(\square \)
Corollary 3
The equation\(A^*(\mu + \delta - \mu z) - z=0\)has only one root in (0, 1) under the sufficient condition\(\delta > 0\).
Proof
Let \(f(z)=-z\) and \(g(z)= A^*(\mu + \delta - \mu z)=H(z)\) as defined in Corollary 1. Consider \(|z|=1- \epsilon \) such that \(\epsilon > 0\) is a very small quantity. Now, \(|f(z)|= 1- \epsilon \) and
Since, \(|g(z)|< |f(z)|\) implies that \(\delta > 0\). Thus applying Rouch\(\acute{e}\)’s theorem it can be concluded that \(A^*(\mu + \delta - \mu z) - z=0\) has only one root in (0, 1). \(\square \)
Rights and permissions
About this article
Cite this article
Kumar, N., Barbhuiya, F.P. & Gupta, U.C. Unified killing mechanism in a single server queue with renewal input. OPSEARCH 57, 246–259 (2020). https://doi.org/10.1007/s12597-019-00408-w
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12597-019-00408-w