QUEUING (WAITING LINE) THEORY Long queues are generally seen in front of railway booking offices, post offices bank counters, bus stations, cinema {\:ket windows ete “Similarly we find automobiles waiting at service ¢ ations, airy ane waiting for landing, patients waiting for doctors etc. Thus queu mat be of persons waiting at a docter’s clinic or at railway booking 0. fice ‘may be of machines waiting to be repaired, or may be of letters arr ving services (known as customers) wait for services or when the service facilities stand idle and wait for customers. Some customers wait when the total number of customers exceeds the number of facilities. Some service facilities stand idle when the number of service facilities ex- ceeds the number of customers. However queues are very common phenomena of modern civilized life. A group of items waiting to receive service is known as wailing line. or queue. A queue is formed when the demand for a service exceeds _ the capacity to provid: that service. a Queuing theory is a quantitative technique which consists in | constructing mathematical models of various types of queuing systems. _ These models can be used for making predictions about how the. system can adjust with the demands on it. Object of Queuing theory. ie Customers wait for service. The time thus lost by them is expen- _ sive. The costs associated with waiting in line are known as waiting " time costs. Similarly if there aré no queues, ‘Service stations will be | idle. Costs associated with seryice or the facility are known as service costs. The object of queuing theory is fo achiove an economic balance between these two types of cos's. That is, study of queuing theory helps in minimising the total waiting and service costs. at the clerk’s table for despatch. Itis formed when the units requiring —# Queuing theory can be used to estimates the different characteristics of the waitingline such as (1) average arrival rate @) average service rate (3) average waiting time (4) average time spent in the system etc. f Application of Queuing theory ‘The study of queuing theory is mainly applied in the following fields. i. Business: Queues are formed every day, in front of banks, super- market, booking offices etc. i 2 Indesteies: In servicing of machines, storage etc. queues are quite > 3. Engineering: In the field of Telephony, electric, computers etc. also queues are frequently formed. _ 4. Transportation: In postal services, airports, harbours, railways etc queues are formed for transportation. 5 Cisher fields: Cinema ticket window, barbershop, restaurants efc. are not able to avoid queues. Im all these fileds queuing theory can be applied to the problems connected with queues. Certain terms commonly used \. Customer: Person waiting in a queue or receiving the service. 2. Server: Person by whom service is rendered. ‘ ae | 4. Time spent by acustomer: Waiting time plus service time. 4, Number of custamers in the system= plus the number of customers bemg serviced. 5. Queue length: Number of customers waiting in the Queue. : & Queui m= System consisting of arrival of customers, waiting in Gosie Gaked up of service, being served and the departure of cus- tomers. Queuing Process ‘A Queuing process is centered around a service system which has one or more service facilities. ‘Customers requiring service are arriving Number of customers in the queue - onditions). Any customer entering the service system joins a queue for service facilities. Customers are selected for service by some rule, known as service discipline. After the service is completed the cus- "tomer leaves the service system. _ Characteristics of Queuing system (Elements of the Quening system). 1) Theinput process orthe arrivals: ‘The input source is finite or iefinit-. ; ‘The input describes the way in which the customers arrive and join the system. Generally customers arrive in a. more or Iss random fashion. ‘The arrival times of customers in a restanrant are cistributed more o- ” Tess randomly arid cannot be predicted. Arrivals may occur at regular _ intervals also. Ina clinic, patients may be given appointments in such i ‘a manner that they arrive at the clinic at specified equal intervals of | time. ‘Arrivals may occur af a constant rate or may be in accordance _ with some probability distribution such as Poissy m distabution or | Normal distribution. “Thus the arrival pattern can b. st be descr? edit _ terms of probabilities. : _ 2) Service mechanism: Service mechanism concerns with the servic ” time and service facilities, Service can either be fixed or distributed iu, be (a~ one queue - one service station (single channel facility) mW) one queue - several service station. © several queues - one service station. (@) many queues and many service stations (Multi channel facilities). 3) Queue discipline (or service discipline): If any of the service facilities '_ are free, the incoming customer is taken into service immediately. If F however, all the servic: facilities are busy, the customers in the queue a may be handled in a number of ways as service facili y becomes free. _ Some of these being: @ ‘The customers are selected for service at random. This is known as the ‘random’ service discipline. This is found in many operational, situations where the customers do not wait in a well } “organized line.K.4 (ii) The customers are assigned priorities. The service facility becoming free, commences service on the customer with the highest priority, Ifthere is more than one customer of the same priority in the queue the service facilities may select a customer from among these either on the ‘fist-come, first served’ or random basis. ” Queue discipline also refers to the manner in which the customers form into queue and the manner in which they behave while being in the quene. _ A customer may decide to wait no matter how long the queue becomes. Some customers may decide not to enter the queue because of its huge length. Some customers after entering the queue and - waiting for some time lose patience and leave the queue. When there are more than one queue the customers may move from one queue to another. i 4} Output of queue: In a single channel facilities the output of the queue does not pose any problem for the customer who leaves after 1% getting service. But it is important when the system is multistage iS we facilities, because the ossibility of a service station break 09 down can have repercussion on the queues. x pots ay AC Limitations of Queuing Theory v sh, Most of the Queuing models are com) lex and cannot be easily 17 ae understood. There is always the element of uncertainty in all queuing situations. There the probability distribution to be applied for arrivals or servicing may not be clearly known. « a2 2. Queue discipline also imposes some limitations. If the assumption iS oe ‘of first come first served is not a true one,” Queuing analysis 40\\0* becomes more complex. 4a yi Jn multichannel queuing, several times the departure from one “ queue forms the arrival for another. This makes the analysis more complex. Role of Queuing theory in management Queuing theory plays a very important role in the management. Decision regarding the amount ‘of required capacity must be made frequently in industry and elsewhere. These decisions are often dif- ficult ones since it is impossible to predict accuratelywhen units will arrive to seek service, how much will be required to provide that service ete. Providing very much service facility would involve excessive costs. Ifenough service facilities are not provided, that will cause long waiting Tines to form. Excessive waiting is also costly. Therefore ultimate goal K.5 of queuing theory is to achieve an economic balance between the cost of sercvice and the cost associated with waiting. Based on probability theory it attempts to minimise the extent and duration of investment cost. Queuing theory gives the estimated average time and intervals under sampling method and helps in taking decision of optimum capacity so that the cost of investment is minimum keeping the amount of queue within tolerance limits. PROBLEMS (M/M/D) and (M/M/K) Formulae 1. W stands for mean arrival rate or Expected arrival rate. 2. ‘w’ stands for mean service rate or Expected service rate. 3. ‘Traffic intensity or probability that the service facility is not idle eS a 4, Probability that service facility is idle or probability that no units are inthe system = Po = 1 -4 , # 6. Probability that the number in the queue being served is atleast ‘n’ 5, Probability for ‘n’ units in the system = Py = Gy ( ce il ey B 7. Average number of units in the system = a P38. Average number of units in the queue (Average length of Queue) : ¥ : ~ EEA 9. Average time a unit spends in the system (waiting time plus service i a eu time) gaa 40, Average time a unit spends waiting in the queue = a Z x HE’ 11. Average length of non empty queue = a i pe Qn. 1 Arepair shop attended by a single machine has an ayerage of four customers an hour who bring small appliances. The mechanic inspects them for defects and quite often can fix them right away or otherwise render a diagnosis. This takes him six minutes, on theK.6 average. Arrivals are Poisson and service time has the exponential distribution. You are required to (@) Find the probability that the shop is empty. (b) Find the probability of finding at least one customer in the shop. (c) Whatis the average number of customers in the system (d) Find the average time spent, including service. Ans: Mean arrival rate, 2 = 4 customers per hour Meanservice rate, “# = 3X 60 =10per hour = 6 (a) Probability of nobody in the system = 1 -4 = Ls (b) Probability of at least one customer in the system = Prob. that the service facility is notidle. = 77 = 4 (oc) Average number of customers in the system nee eae i ed T4048 (d) Average time spent in the system A is F hours or 10 minutes. poi 10-4 Qn, 2 Ina railway marshalling yard, goods trains arrive at arate of 30 trains per day. Assuming ‘that the inter-arrival time follows an exponential distribution and the service time distribution is also exponential with an average 36 minutes. Calculate the following: (a) Average length of queue (b) the probability that the queue size exceeds 10. Ans: i LCOS UL Ui i Mean arrival rate, 2 = Ci ae trains / minute Mean service rate, # = # trains / minute., . = #8 2 Atrains (a) The average length of queue al 7 Z- ale (b) Prob. for queue size exceeds 10 = P(n= 11) = @" = e" Qn. 3. Customers arrive at a booking office window, being manned by a single individual at the rate of 25 per hour. Time required to serve a customer has exponential distribution with a mean of 120 seconds. Find the average waiting time of a customer. Ans: A = 25 per hours Kay x60 ‘ ileago wo 30 per hour Average waiting 7 _ 4 25 f TEES ANE RCS MST SEY 5 *, Gace 2 pG-D ~ 0G0=B) 7 g hour = 10 minutes Qn. 4, A TV repairman finds that the ti : _ EV repairm: ime spent on his jobs has an oe distribution with mean 30 minutes. If he repairs sets in o in which they came in, and if the arrival of sets is ap- pee! Peissou ne ae average rate of 10 per 8 hour day, what repairman’s expected idle time each day. Hoy j of the average set just brought in ? ee Ans: He So i : flere A axe ~ ag Sets/minute and p = a sets / minute Prob. that there is no unit in the system = Prob. that facility may be ‘idle =P) = 1-4= 1-3 =3 ~. Repairman’s expected idle time in 8 hour day = 2 x 8 =3hr and average number of jobs waiting = A. =12 jobs. pe 3 am 5 is belt snapping for conyeyers in open cast mine occur at the : of2 per shift. There is only one hot plate available for vulcanising and it can vulcanise on an average 5 bults snap per shift. 1. What is the probability that when a boll i Pe anit a iy it snaps, the hot plate is _ 2. Whatis the average number in the system ? _ 3. Whatis the waiting time of an arrival ? ! 4. What is the average waiting time plus vulcanising time ? ) ‘Ans: A = 2beltspershift; y = 5 belts per shift. Prob. that there is idling = Py) = 1-4 = 1-2 # Ss tl ule Average number in the system = win eae wd Ue Ae Aula 2a By anes Gz) S622), 5 Average waiting time plus vulcanising time = Average waiting 1 iL 1 time in the system = —~ = —~ = = y Hea 5-2 163 Average waiting time of an arrival =K.8 Ex. 6 The mean rate of arrival of planes at an airport during the peak period is 20 per hour, and the actual number of arrivals in any hour follows a Poisson distribution. The airport can land 60 planes per hour on an average in good weather and 30 planes per hour in bad. sveather, but the actual number landed in any hour follows a Poisson distribution with these respective averages. When there is congestion, the planes are forced to fly over the field in the stack awaiting the landing of other planes that arrived earlier. \ () How many planes would be flying over the field in the stack on an average in good weather and in bad weather ? (id) How long a plane would be in the stack and the process of landing in good and in bad weather ? Ans: Inthis problem, we are given A = 20planes/ hour ad ds & a planes/hour in good weather 30 planes/hour in bad weather (i) Average number of planes in the queue. RR C / {60 (50-20)] = ; Ga good weather) = (uA) | 20°7 [30 G0-20)] = 4 Gnbad weather) (ii) Average waiting time in the system 1 [17 0-20) = A hrs. (in good weather) eh (G04 20); = hrs. (in bad weather) Ex. 7 Customer arrive at a one window drive in bank according to Poisson distribution with mean 10 per hour. Service time per cus- tomer is exponential with mean 5 minutes. The space in front of the windows, including that for the serviced car can accommodate a maximum of 3 cars. Other cars wait outside this space. (a) What is the probability that an arriving customer can drive directly to the space in front of the window ? (b) What is the probability that arriving customer will have to wait outside the indicated space ? (©) How long is an arriving customer expected before starting ryice ? Ans; Here A = 10 per hour, # = 60/5 = 12 per hour. {a) The probability that an arriving customer can drive directly to the space in front of the window. =. Prob that 2 or less in the system = Po +P, + P2 Average waiting time for depositors = K.9 ‘ E OR eh a a ge a-t) +h 0-f) + GP G-p = 0-2) het Qh. =a-2) pres Qi = 02 : _©) The probability that an atriving custome, has to walt outside the indicated space. = Prob. that more than 3 in the systent” he = @f = GS" = 08 (©) Average waiting time of a customer in quene i : HG@-A) 1242-19) Ex. 8 Abank hag two tellers working on savic: teller handles withdrawals only. The secouc handles de, asit. only. It has ound that the service time distributions for otk deposits and withdrawal are exponential with mean service time = minutes per customer. Depositors are found to arrive in a Poisson fashion throughout the day with imean axrival rate 16 per hour. Withdrawals also arrive in a Poissom fashion. with mean arrival rate 14 per hour. ‘What would be the effect on the average waiting time for depositors and withdrawers 2 Ans: Mean arrival rate of depositors = 4, = 16-per he. Mean arrival rate of withdrawers =a, = idperhr. Mean service rate for hoth tellers =e = 5 x 60 = 20perbr. = 0417 acounts, Th firs. deena ga ees) F@-h) (7000-16) = hour = 12 minutes. and Average waiting time for withdrawers = = A a #@-h) 0 Q0-14) = 4S hour = 7 minutes. Ex. 9 A repairman is to be hired to repair machines which breal- down at an average rate of 6 per hour, The bveakdown follows Poisson distribution. The productive time of a machine is considered to cost Rs. 20 per hour. The repaixmen, Mx. X and Mr. ¥ have tcenK.16 i. inferviewed for this purpose. Mr. X chargesRs. 10 per hour and he ‘services breakdown machines at the rate of 8 per hour. Mr. Y demands Rs. 14 per hour and he services af an average rate of 12 machines per hour. Which repairman should be hired? oh howe shrift per day}. Ans: 4 ‘= 6perchour. x= per hour for X and 12 per hour for Y. Mr. & 2 ee ee ee eee eee 7 ae 2 =. Machine hours lostinShour = 3x 8 = 24hrs. Cast of the productive time ofa machine = 20Rs. <. The cost for24¢hrs = 24x 20 = 480. Fora day (8 hrs), the charge for therepair man = 8 x 10 = 80Rs. <. Fotalcost = 480 + 80 = S60Rs. : Mr. ¥ Average number of machines = mE = <=. Machine hours lestinShour = 1x 8 = 8hrs. Total productive cost for these 8hrs = 8 x 20 = 160Rs. Charge for repairing per hour = 14Rs_ Charges for the whole day = 8x 14 = 112Rs_ <. Totalcost = 112 + 160 = 272Rs. : ‘Therefore repairman Y should be preferred. Ex. 10 Arrivals of machines ata tool crib are considered to be Poisson distribution at an average rate of 6 per hour. The service time at the tool crib is exponentially distributed with an average of 3 minutes. (1) What is the probability that a machine arriving at the tool crib will have to wait? (2) What is the average number of machinists at the fool crib ?. @) The company will install a second tool crib when convinced that a machinist would have to wait atleast three minutes before being served. By how much the flow of machinists to the tool crib must increase, to justify the addition of a second tool crib ? Ans: A = 6perhour = Z per minute = 20 per hour. =3 fagat ‘ (1) Prob that the too! ccib will have to wait = Prob that tool eribis notidie = 4 = inthe queue = 3min = 05 hour. 2 a, = 05% 20 (20-2) 2 4 =1@0-2) = At+4,= 2 = 24, = 20 : 2. Ay = 10 per hour So arrival should be 10 pex hour + Increase im arrivalraté = 10~6 = 4 per bour. =. There should be an increase by, pes owt asthe ston of a second tool cxib. -Ex. i Arrivals at a telephone booth are considered to be Poisson, with an average time of 1) minutes between one arrival and the next. The length of a phone call assumed to be distributed exponentially, with mean 3 minutes. Find the following. (a) What is the probability that a person arriving at the booth will have to wait ? (L) What is the average length of the quene that form from time to tirne ? (c) The telephone company will install a second booth when c: nyinced that an arrival would expect to have to wait atleast three minutes for the phone. By how much must the flow of arrivals be increased in order to justify a second booth? (a) Find the average number of units in the system {e) Estinafe fraction ofa day that the phone will be in use. Ans: Here, mean arrival rate, A= # per minute = 6perhour and mean service rate, 2 = 3permin.= 20 per hour (a) Prob. [arrival will have to wait] = Prob that facilityis not idie = 2 - £ = 03Ria (c) See the last question (same answer) {@ Seethelast question (same answer) eran Oki that the will be in : {e) Fraction of a day phone Die ioe Goes a = Prob, that the phone is notidle = 7 ' 4 ire the service 12 Onanaverage 96 patients per 24 hour day require thes ere emergency clinic. Also om an average, a patient requires 10 minutes of active attention, ‘Assume that the facility can handle only one emergency at a time. Suppose that it costs the clinic Rs. 100 per patient treated to obtain an average servicing time of 10 patients and Twat ench minute of decrease in this average time would cost Rs. u per patient treated. How much would have to be budgeted hy the clinic to decrease the average size of the queue from 4j patients to; patient. Ans: Arrivalratea = % = 4perbour. ‘ Service rate = # = % permin = 6 per hour. Average number of patients in the queue 22 fi St patients: “pe-” 66D : ‘We want to reduce this to #’ be the new service rate. 1 ww 2 i a EASA OH = " = Sper hour. Be 4), 2 E Service rate is 8 for reducing the average size of queve ‘Time required for one service = $ hour = 75 minutes 10-75 eB i time of treatment a 2. Decrease in the average peters Cost of the clinic before decrease: = 100 Rs. Cost of the clinic after decrease: = 100 425% 10)= 125Rs. “The budget should be Rs, 125 per patient. it ives for cashing Ex. 13 Ina Bankevery 15 minutes one customer arrives : the cheque. The staff in the only payment counter takes 10 ‘minutes for serving a customer on an average. State suitable assumptions and find (1) the average Queue length (2) increase in the arrival K.B ' rate in order to justify a second counter (when the waiting time of a ‘customer is atleast 15 minutes the management will increase one or more counter). : Ans: Assumptions: (a) Arrival pattern follows Poisson distribution (b) Service time follows an exponential distribution. Arrival rate = 4 = 2 per minute = 4 per hour Servicerate = #@ = + per minute = 6 per hour @ Average Queue et (Average number of units in the Queue) ~ = ear coo aA (2) The second counter is justified when the average waiting time of an arrival in the queue should be atleast 15 minutes: Let the new arrival rate for this be 2’ AI hs B@-*) 6 2 604! = 154 (u-A’) “ 604" = 15x 6(6-A') 2 604" = 540-902" <. 1904' = 540 ord! = 3 = 3.6 For increasing the number of counters minimum arrival rate = A! = 3.6 per hour. Present arrival rate = 4 = 4perhour So even now a second counter is justified. . REVISIONARY EXERCISES What do you understand by queue ? Give some important applica- tions of queuing theory ? 2. Give the essential characteristics of the queuing process. 3. What do you understand by (a) queuing model (b) queue dis- _ cipline. . ) 4, Explain the constituents of a queuing model. 5. Explain the object of Queuing theory. 4