Simulation of a Queuing System Case Study: Call Center Service in Organization A. Rutcharin Kullachart, Kasetsart University, Thailand

Transcription

1 Siulation of a Queuing Syste Case Study: Call Center Service in Organization A. Rutcharin Kullachart, Kasetsart University, Thailand Asian Conference on Psychology and Behavioral Sciences 2015 Official Conference Proceedings Abstract The objectives of the study is to fine the queuing syste of the call center service in Organization A by using the siulation situation. The siulation situation ain purpose is to fine the queuing characteristics and offer the alternative way to reduce the custoers queuing tie. This study use the ARENA progra to siulated situation to learn about syste efficiency under the uncertainty of the tie to spend on service. And the uncertainty of the arrival of custoers. Assues that arrival tie follow a Poisson probability distribution and the service ties are distributed exponentially. In this study the queuing characteristics of the call center service in Organization A is analyzed by ultiple-channel queuing odel to fine the average nuber of custoers in the syste, the average nuber of custoers in line waiting for service, The average tie a custoer spends in waiting line, the average tie a custoer spends in the queue waiting for service and Utilization rate. This study use ARENA software to copute the perforance easure of Multi-server queuing syste at call center service in organization A using arrival rate ( ) custoers/hr, service rate ( ) 25 custoers/hr and nuber of service () 27. The study results on efficiency of the queuing syste of call center of Organization A revealed that even though there is a change in nuber of outsourcing agents, the efficiency values of queuing syste are still close to the current situation where 21 outsource agents are eployed. Keywords: The queuing syste, Multiple-channel queuing odel, ARENA progra iafor The International Acadeic Foru

2 Introduction Probles on operation of call center which affect service rendering efficiency are uncertainty of incoing calls fro the custoers (arriving custoers) and uncertainty on aount of tie each telephone operator rendered service. For instance in soe period of tie the nuber of arriving custoers exceeds the nuber of telephone operators or soe telephone operator takes too uch tie to provide service to each custoer due to such operator does not have enough skills. Consequently, soe custoers call ay not be answered iediately and they have to wait before the operator pick up their call. Therefore soe custoers shall not wait as it akes the have waiting cost, so they choose to purchase products fro other copany instead. On the other hand, if there are too any operators, service cost shall be higher. The above probles ade the researcher interested to conduct an analysis on queuing syste of call center of one organization. The objectives are to find guidelines to develop efficiency of services, to ake service user have faster service, to balance waiting cost and service cost, to create business operation worthiness of the organization and to axiize satisfaction of the service users. Objectives of the Project 1. To conduct a study on queuing syste on service rendering in call center of Organization A 2. To conduct a study on queuing syste odel in call center of Organization A. Related Theories Siulation of Proble Siulation of proble is a odel design process of real syste, then conduct experient to learn behavior of real syste under the specified requireents to assess operation of the syste and the experiental result shall be analyzed before it can be used to solve probles in the actual circustance (Shannon, 1975, referred to in Roongrat Pisuchpen, 2008). As the current industries always have coplexity with variable factors involved. For exaple, in service industry, the variable factors are uncertainty of arriving custoers and uncertainly of service rendering tie and etc. Therefore, it is quite difficult to anage the works and allocate resources to ake the achieve the objectives of such industry by using arithetic odel to solve the proble. Hence, creation of odel by using coputer progra is one of the options to study behavior of the syste to find guidelines to solve the proble. In addition, when using coputer progra to create odel, a change of odel to create guidelines to iprove the syste to ake it have better efficiency and effectiveness without any interference of real syste can be done quite conveniently (Roongrat Pisuchpen, 2008).

3 Queuing Syste Queuing Theory has been initiated and developed by A. K. Erlang, a Danish engineer, in Queuing occurred because deand for services exceeds capability of service rendering, therefore it is very iportant to have adequate service unit to render services. In order to be able to provide service adequately with the deand, it is necessary to know the aount of arriving custoers and when as well as tie used for rendering service to each custoer. If there are too sall aount of service units, queuing shall occur, which is deeed as one of the loss of expense probles and it can also ake the business lose the custoer. On the contrary, if there are too any service units, expenses incurred shall be ore than necessary, for instance service unit shall have too uch idle ties (Sutthia Chunanvej, 2011). Procedures to Conduct Research Collect inforation on nuber of receiving custoers, tie used for providing service of operator in call center, period of tie where there are a lot of arriving custoers ( ), fro database of Organization A. Inforation gained fro database shall be used to create odel to analyze queuing circustance at present by using coputer progra, such as service usage rate and service rendering rate of the telephone operator in order to find average service user in queuing syste (L), average nuber of service user in the queue (L q ), average tie each service user is in the queuing syste (W), average tie each service user is in the queue (W q ), probability of syste availability (P o ). After that, odel shall be crated to test the concept to iprove efficiency of the queuing operation syste. Research Methodology 1. Copile inforation on tie used for service and copletion tie for providing service of the operator fro call center database of Organization A. 2. Study and analyze inforation to specify the appropriated queuing odel as M/M/ queuing odel under assuption that usage of service user is on rando basis with Poisson Arrival Process Distribution, tie used for providing service is on rando basis with Exponential Service Tie Distribution, there are ore than 1 service unit ( Unit), 1 queuing and with one step service rendering, unliited nuber of service usage population, average service usage rate () is less than average service rendering rate ultiplied with nuber of service unit (), average service rendering rate of service unit is equal.

4 For the odel the probability of having n custoers in the syste is given by: 1. The probability that there are zero custoer in the syste: ) ( ) (!! 1 n n n n P 2. The average nuber of custoers in the syste: ) 1)!( ( ) / ( P L 3. The average tie a custoer spends in the syste: L P W + 1 ) 1)!( ( ) / ( The average nuber of custoers in line waiting for service: L L q 5. The average tie a custoer spends in the queue waiting for service: q q L W W 1 6. Utilization rate: ρ Where, the arrival rate of custoer per unit tie, the service rate per unit tie, the nuber of servers, P 0 the probability that there are no custoer in the syste, L q the average nuber of custoers in line waiting for service, L the average nuber of custoers in the syste, W q the average tie a custoer spends in the queue waiting for service, W the average tie a custoer spends in the syste, ρ Utilization rate. Details of Current Proble Call Center of Organization A provides service 24 hours. Period of to hrs. is the tie that the call center has the largest aount of arriving custoers and outsourcing agents are hired to support the work in such period.

5 At present, during period of hrs., there are 6 peranent agents of Organization A and 21 outsourcing agents provide services on such period. Average arriving custoer is inutes and service rendered is one person each and there is no liited nuber of axiu service users. In order to enhance efficiency of operation, the researcher has conducted a study on balance between nuber of telephone operators and service usage rate of custoer. Create Model by Using Coputer Progra This study use ARENA software to copute the perforance easure of Multi-server queuing syste at call center Service in organization A using arrival rate ( ) custoers/hr, service rate ( ) 25 custoers/hr and nuber of service () 27. Model is starting fro Create Module, naed Custoer Arrival into the syste on randoly basis (Expo) with average value of inute and service used is 1 person at a tie, without liiting the axiu service users. Next procedures after the custoer is in the syste are to assign odule naed Count Tier to specify qualifications to the custoer to record starting tie when the custoer is in the syste, assign qualification naed TieEnter to TNOW to collect current tie when the custoer enters into Count Tier Module to calculate tie that custoer used in the syste at a later tie by using Record Module. After the custoer enters into the syste, there shall be an Interactive Voice Response (IVR) via Process Module naed IVR. Define operation procedure as Delay with unliited resources, so it is not necessary to reserve the resources, so no queue is occurred. Tie used for perforing activity is fixed (Constant) at seconds. Next, the custoer uses service with Agents via Process Module naed Agents Process. Define operation procedure as Seize Delay Release to reserve the existing 27 agents. After the activity is copleted, the agent shall be set to available to provide service to the next custoer. Use Set resource naed Agent and Delay Type is Expression with 27 ebers in the Set. Type of resources is specified to be non-constant production capacity based on schedule. Working tie of agent is specified to be 8 hours with orning interval for 15 inutes, lunch break for 40 inutes, afternoon break for 15 inutes. Schedule Rule is specified to be as Wait. When it is the tie to break, but if the service rendered to custoer by the telephone operator has not yet copleted, such operator shall keep on providing service until it is copleted and shall take a break after that. When such operator resues the work, starting tie shall be extended, so total period of tie in aintaining workforce of agent shall be the equally sae. After that Record Model shall be created as Tie Interval to record tie and nuber of custoers who finish service usage with the Agent, and custoers exit fro the syste.

6 Nuber of agents 8.00 a a a a a a a p p p p p p p p p.. 3 PA 3 PA 7 OA 7 OA 7 OA Nuber of agents: *PAPeranence Agents,*OAOutsource Agents Figure 1: Working Tie Table of the Agents (during period of hrs.) Nuber of Agents Morning break (15 inutes) Lunch break (40 inutes) Afternoon break (15 inutes) 3 Agents a a a p p p.. 3 Agents a a p p p p.. 7 Outsource agents 7 Outsource agents 7 Outsource agents 9.45 a a p p p p a a a p p p a a p p p p.. Nuber of agents: *PAPeranence Agents,*OAOutsource Agents Figure 2: Peranence Agents and Outsource Agents break Schedule Suary of Current Situation of the Call Center of Organization A by using ARENA Siulation Progra is as per Figure 3. Based on the study, the current situation of the Call Center of Organization A is as follows: 1. The period that call center has the largest aount of arriving custoers is fro hours and there are totally 27 agents, coprising of 6 peranent agents and 21 outsource agents, provide services at that tie. 2. The average arrival rate of custoer custoers/hour. 3. The average service tie of each agent 25 persons/hour. 4. The average nuber of custoers in the syste (L) 5.91 persons. 5. The average nuber of custoers in line waiting for service (Lq) 0.01 person. 6. The average tie a custoer spends in the syste (W) 3 inutes. 7. The average tie a custoer spends in the queue waiting for service (Wq) 0 inute. 8. The average utilization rate of 27 agents 21%.

7 Nuber of agents W in inutes Wq in inutes L Lq Utilization rate (%) 6PA+21OA (Present Situation) Nuber of agents: *PAPeranence Agents,*OAOutsource Agents Figure 3. Perforance Measure of Call Center Queuing Model at Organization A (Present situation) Efficiency Iproveent Guideline: Siulate service provision of a call center during hours by specifying that the average arrival rate of custoer is fixed at persons/hour, the average service tie of each agent is fixed at 25 persons/hour and working schedule of each peranent agent and outsource agent is fixed. After that, conduct a test on efficiency coparison of queuing syste in case of a change in nuber of the outsource agents and the current situation, result gained is as per Figure 4. Nuber of agents W in inutes Wq in inutes L Lq Utilization rate (%) 6PA+21OA (Present Situation) 6 PA + 0 OA PA + 1 OA PA + 2 OA PA + 3 OA PA + 4 OA PA + 5 OA PA + 6 OA PA + 7 OA PA + 8 OA PA + 9 OA PA + 10 OA PA + 11 OA PA + 12 OA PA + 13 OA PA + 14 OA PA + 15 OA PA + 16 OA PA + 17 OA PA + 18 OA PA + 19 OA PA + 20 OA

8 Nuber of agents: *PAPeranence Agents,*OAOutsource Agents Figure 4: Efficiency coparison of queuing syste between the current situationand in case of a change in nuber of the outsource agent Data in Figure 4 can easure efficiency of queuing syste fro the siulation by siulating a change in nuber of the outsource agent and it was found that if there is a change in nuber of outsource agents, the result is as follows: 1. The average tie a custoer spends in the syste (W) is still 3 inutes which equals to the current situation of the call center. 2. The average tie a custoer spends in the queue waiting for service (Wq) is still 0 inute which equals to the current situation of the call center. 3. The average nuber of custoers in the syste (L) is between persons, which is closed to the current situation which is 5.91 persons. 4. The average nuber of custoers in line waiting for service (Lq) is between person, which is closed to the current situation which is 0.01 person. 5. The average utilization rate is between %, which is closed to the current situation which is 21%, result gained is as per Figure 5-9. Nuber of agents: *PAPeranence Agents,*OAOutsource Agents Figure 5: Coparison of the average tie a custoer spends in the syste (W) between the current situation and in case of a change in nuber of the outsource agent Fro Figure 5: When coparing the average tie a custoer spends in the syste (W) between the current situation and in case of a change in nuber of outsource agents, it was found that the efficiency of queuing syste on the average tie a custoer spends in the syste (W) during a change in nuber of outsource agents equals to the current situation which is 3 inutes.

9 Nuber of agents: *PAPeranence Agents,*OAOutsource Agents Figure 6: Coparison of the average tie a custoer spends in the queue waiting for service (Wq) between the current situation and in case of a change in nuber of the outsource agent. Fro Figure 6: When coparing the average tie a custoer spends in the queue waiting for service (Wq) between the current situation and in case of a change in nuber of outsource agents, it was found that the efficiency of queuing syste on the average tie a custoer spends in the queue waiting for service (Wq) during a change in nuber of outsource agents equals to the current situation which is 0 inute. Nuber of agents: *PAPeranence Agents,*OAOutsource Agents

10 Figure 7: Coparison of the average nuber of custoers in the syste (L) between the current situation and in case of a change in nuber of the outsource agent Fro Figure 7: When coparing the average nuber of custoers in the syste (L) between the current situation and in case of a change in nuber of the outsource agents, it was found that the efficiency of queuing syste on the average nuber of custoers in the syste (L) during a change in nuber of outsource agents is between persons which is closed to the current situation which is 5.91 persons. Nuber of agents: *PAPeranence Agents,*OAOutsource Agents Figure 8: Coparison of the average nuber of custoers in line waiting for service (Lq) between the current situation and in case of a change in nuber of the outsource agent Fro Figure 8: When coparing the average nuber of custoers in line waiting for service (Lq) between the current situation and in case of a change in nuber of the outsource agents, it was found that the efficiency of queuing syste on the average nuber of custoers in line waiting for service (Lq) during a change in nuber of outsource agents is between person which is closed to the current situation which is 0.01 person.

11 Nuber of agents: *PAPeranence Agents,*OAOutsource Agents Figure 9: Coparison of the average utilization rate between the current situation and in case of a change in nuber of the outsource agent Fro Figure 9: When coparing the average utilization rate between the current situation and in case of a change in nuber of the outsource agents, it was found that the efficiency of queuing syste on the average utilization rate during a change in nuber of outsource agents is between % which is closed to the current situation which is 21%. Conclusion The siulation study results on efficiency of the queuing syste of call center of Organization A by using ARENA Siulation Progra revealed that even though there is a change in nuber of outsourcing agents, the efficiency values of queuing syste are still close to the current situation where 21 outsource agents are eployed. This can signify that if the nuber of the outsource agents are reduced by 6 persons, the efficiency values of queuing syste are still close to the current situation. Therefore, for optiu cost efficiency on eployent, the anageent of call center of Organization A should eploy only 15 outsource agents and should assign 6 reaining outsource agents to perfor other works or delay eployent of the outsource agents in the long ter.

12 References Chochai, Y., & Chancharatsuk, U. (2013). Perforance Iproveent of Queuing Syste by using Siulation. A Case study: The out Patient Departent, Lalukka Hospital. (In Thai). Ladkrabang Engineering Journal, 30(1): 43-48, Retrieved Noveber 15, 2104, for Duong, T., Huu, N. D., & Nguyen, T. (2013). Adiabatic Markov Decision Process with Application to Queuing Systes. Proceeding of Inforation Sciences and Systes Annual Conference, 47, 1-6, Retrieved Deceber 2, 2012, for Guruurthi, G., & Benjaafar, S. (2004). Modeling and Analysis of Flexible Queuing Syste. Journal of Naval Research Logistics, 51(5), , Retrieved February 15, 2014, for Kebe, M. M, Onah, E. S., & Lorkegh, S. (2012). A Study of Waiting and Service Costs of Multi-Server Queuing Model In Specialist Hospital. International Journal of Scientific & Technology Research, 1(8): 19-23, Retrieved Noveber 15, 2104, for A-Multi-Server-Queuing-Model-In-A-Specialist-Hospital.pdf Olaniyi, T.A. (2004). An Appraisal of Cost of Queuing in Nigerian Banking Sector: A Case Study of First Bank of Nigerian Plc, llorin. Journal of Business & Social Sciences. 9(1): , Retrieved Noveber 15, 2104, for Patosuttirut, V., Piriyakul, M. (2013). Decision Support Syste on the Adinistration of Data Service Agents at Social Security Call Center # (In Thai). Manageent Journal Faculty of Manageent Science Lapang Rajabhat University, 6(1): 55-62, Retrieved Noveber 15, 2104, for Pisuchpen, R. (2010). Siulation with Arena Solutions Manual (Rev.ed). (In Thai). Bangkok : SE-Education. Render, B., Stair, R., & Hanna, E. M. (2012). Quantitative Analysis for Manageent. 11 th ed. New Jersey : Prentice Hall International, Inc. Rosenquist, C.J. (1987) Queuing Analysis: A Useful planning and Manageent technique for radiology. Journal of Medical Syste, 11(6): , Retrieved Noveber 20, 2104, for Rungpeng, P. (2013). Siulation of queuing systes for outpatient service: a case study of the internal edicine at Phatthalung hospital. (In Thai). Veridian E-Journal, 6(3): , Retrieved Deceber 15, 2104, for

13 S. K. Dhar, Tanzina Rahan. (2013). Case Study for Bank ATM Queuing Model. Journal of Matheatics, 7(1): 01-05, Retrieved Noveber 20, 2104, for Taha, Handy A. (2007). Operation research: an introduction. 8 th ed. New Jersey : Pearson Education, Inc. Vanichbuncha, K. (2010). Quantitative Analysis.(In Thai). 1st ed. Bangkok : Chulalongkorn Business school. Contact eail: aa_eaw82@hotail.co