DEPARTURE LOUNGE SIZING AND OPTIMAL SEATING CAPACITY FOR A GIVEN DIRCRAPT/PLIGHT MEX ~ (4) Single Gate. (4S) Several cater S.C. Wirasinghe* M. Shohatet® Department of Civil Engineering ‘The University of calgary 2500 university Drive Nai calgary, Alberta, Canada man ind the efze and seating capacity of an alzport departure lounge is considered. Slieratt/flight type and subject to the constraint that boarding comences Ge close ae possible to the door closing time. The lounge area for pecsengers is a function of the number of seats and standing-spaces since The area for a seat ie approximately S04 greater than the area for a Standing space, An optimal number of seats and a related lounge size existe for a given aircraft/#light type, if a penalty is attached to Conpsisory passenger standing time in comparison to tine spent aiteing. She cbjective of the optinisation is the minimization of the lounge cost plus the total penalty for standing passengers. An approxinate closed form olution iz obtained when the pastenger arrival rate into the lomge is fron a security. check at the lounge entrance, The sensitivity of the Sounge size to the penalty per passenger per hour se investigated, The analyeis is extended to cover the cane of @ mix of alrcraft/flight types Using a gate, The size of a lounge that is shared by several gates us also Gnelyseds the necesoary Lounge capacity can be reduced by up to S08 by combining several Lounge + professor of Civil Engineering, ‘+ Research Associate (part-time)1.0 mwmooucrt0N ‘Air terminal design requires the analysis of passenger and bag through the terminal, ‘the proper design of passenger space in a terninal should eliminate unacceptable waiting tines and queue length at the service components within the airport, provide sufficient space in the waiting areas to accomodate the waiting passengers at an acceptable level of service and on the other hand avoid unnecessary expen 1 due to overdesicn. ‘the design of departure lounges is considered in this paper. paullin and toronjefé® siustrated the queueing process at a departure lounge using deterministic queveing theory (Pig. 1]. Given the curve F(t) representing the cumulative passenger arrivals to the lounge for a chosen aireratt/ELight combination, and the curve G(t) representing the cuntative peesenger departures at the capacity rate, the maxim passenger ‘accumalation Q occurs in most cases at tine t, when boarding comences. ‘the tounge is designed to hold Q passengers. ‘Tbe maximum accumlation could sccur after ty in the unlikely event that the arrival rate exceeds the departure rate for a significant period of tine. This possibility is neglected in the following analy: : (Fa) enzawt? proposed that the lounge be designed for ..."its peak 15 minute average occupancy throughout the plansing day ..." as determined from 9 aimelotion wodel, ‘Transport Canada? recomended that lounges be desicned to accomodate 90% (charters) er 80% (non-charters) of the pascenger cad for the largest aircraft. tart recomended using 85% of the nunber of passengers on a chosen aircraft/flight. 1.2 Roarding Tine Por a given cumistive passenger arrival curve, F(t), the time at vhich boarding comences essentially determines the lounge capscity in ters of as ‘the earlier that botrding is begun, the snaller the needed ges capacity. Fer reasons of comfort and safety, most passengers would prefer to wait in a lounge whore movenant is relatively unrestricted rather eenFLOW TO LOUNGE CUMULATIVE PASSENGER FLOWS ie = aa PASSENGER ACCUMULATION THE sypical Boarding Sequence (Paullin and #oronjett") Figure 1to sit dn an aircraft. Further, it is Likely to be cheaper to rent lomge space rather than aircraft space to store passengers waiting for departere. ‘the conclusion {s that boarding should commence as close az possible to the scheduled aircraft door-closure tine. 1.2 Chotee of Fie) 1A complication arises with P(t) being different for various fight typer (e.g. charter or scheduled, intercontinental or local). thus the F(t) ‘or fs type of flight that causes people to arrive relatively early may have to be used even if the aircraft ie emailer. Referring to Fig. 2, consider the arrival curves F, (0), {= 1,2 for two fights with the Joad for atzeratt 1, Py (ty) Being larger than the load, Fo(ty), for aircraft 2. The boarding crates, bj, are such that by > bj. If boarding Le started az late as poussbie at tines ty, and tg), the naxinun number of passengers stored in the lounge for the second flight, 9(2), is larger than the equivatent number, 91), for the first flight. It is Likely that aircraft 1 is the aarser fraction of passengers arrive earlier for that f1ight. Given @ ined Lounge capacity of Q(2), aireratt 2 would have to be boarded ster ‘Te arrivals to the lounge can thenselves be the output from several servers auch ae customs, Samigration and security-check that are commen to many flights. While the data reported in the Literature always show the ‘typical s-shape for the curve P(e) representing cumulative arrivals to & tounge for a flight, the arrivals to a single lounge that 1s the output from a common-server servicing several flights can take different shay shom in Fig. 3. The arrivals for two flights 1, 2 are processed for sity by a common server at a rate $_. The resulting cumative departure are shova by the dashed curves. The curves for departures fron curity check for flights 1, 2 are drawn using a nethod proposed by ewe12® for work-quet The curve for departures from security check for flight 2, which is the arcivals curve to the lounge does not have the‘CUMULATIVE PASSENGER FLOWS aca) aun Yeo "eo te 1) TIME Arrival and Departure of Passengers for two DAE nt Aireratt/#light Types.‘CUMULATIVE, PASSENGER FLOWS FLIGHT | ‘COMBINED, ARRIVALS DEPARTURES TO ‘COMBINED DEPARTURES ® DEPARTURES TO FLIGHT 2 Departures From a Comon-Server Security Check.‘the typieal § shape. A lounge designed on the basis of the dashed curve could be smaller. However, if Light 2 took place in the absence of flight 1, & Lounge deeigned on the basis of the dashed curve could be Anoufficient. Thus it is better to ignore the output from comon-servers when estimating the lounge capacity. tlowever, a lounge with a dedicated security check at the entrance can be designed for the output from the security check. (Fie 3) 1,3 Lounge Size ~ Most references (Table 1) recommend that the lounge size be obtained by multiplying the nurber of passengers estinated as described above, by a ‘xed area per passenger. ‘The suggested value for the area/passenger ranges from 0.25 to 2.0 meter*, The range suggested by Hanzawi” is based on variations in the "level-of-service" provided. ‘eanaport Canada? and Hart recognized thet the area per seated passenger Le larger than the area per standing passenger. Transport Canada recomended that seate be made available to 508 of the passengers and Hart appears to be making a similar assumption. The incresse in the area/ passenger is in the range 0.5 to 0.6 neter*. source rea/Passenger (nt) Paulin and Horonjett aa Horonseee™ 2.0 2 ‘reaneport Canada’ 1.0 (standing? 115 (eeateal aantord® olssat stanzas? Area per Passenger for tounge Sizing ‘TableSince the comfort provided to passengers is significantly affected by the avalisbility of a seat, and since the initial cost of the lounge is snoreased by over $1000 per seat due to the increase in necessary area, the price of the chair and installation cost, the number of seats in @ lounge designed for a given mumber of people should be chosen carefuliy. ‘me dtscontort caused by the Lack of a seat is explicitly included in the following analyses by introducing a penalty for each unit of standing tine spent by a passenger. The tradeoff between the cost of providing seats and ‘the penalty paid by standing passengers is exploited to cbtain an optim muaber of seats. Te As possible under certain airport geometries to share a lounge atong ‘several gates. A method for estinating the size of a shared lounge is e180 presented in thie paper. 2.0 OPFIWAL LOUNGE SIZE (Unique Arrivals Curve and Departure Rate) First we consider a lounge designed for a given cunslative passenger fecival curve P(0) and a fined hoarding rite, 5. This ensenttatly means that all flights from the lounge (gtte) are sinilar in tema of fist type, aircraft. slze, load factor and boarding ate. tat the boarding ible at tine {Fi9. 41. (Fig. 4) the minim lounge capacity te then fined at Q, the paanenger sccumasion at ty. The Lounge area, A, for passenger storage under normal conditions 4s given by iI - 81 o 6 = the number of seats provided (50), my) m, © aren per passenger, per teat and per standing-rpace respectively, with my > my.2E a Light ss delayed, the Lounge should be able to accommodate a fall plane-load, ¢, of passengers, at a lover level of service (Less ar oe whore tn, = mininun area per standing passenger. 1 appears from Table 2 that 4y 48 in the range 0.35 to 0.60 meter’. ‘Since a seated space Ls about 50¥ larger than a standing space, the nunber lof seats influences the lounge size and cost significantly, and should be selected with care, 2.1 chotee of § TE 5 seate (5 9] are provided in the lounge, they could all be occupied by time €,[Pig, 4] and all later arrivais my have to stand. The total delay to passengers 1s the ares between the cumulative arrival and departure curves and it is independent of § and the queve discipline. It Le not possible to estinate individual waiting tines of passe ers since boarding is uovally by aircraft gest location and not on a flzst-in, fieat-out (e170) basis. she fact that acne people who were previeusiy required to stand have the option to sit, vhen sone of the people who were seated start boarding, is ot iiiustrated in Fig. 4, and it would opeear at first glance thst none of the seats are occupied after £. This i of course not trie since some seats will become available after tine &. (Fig. 5) in Fig. 5 = curve 1, vhich reprotante the cumlstive miaber of stanting pessengers, it is assumed that « seat {5 always occopied az long a= Shere tre sufficient people in the lounge, and that several people will ute ao‘CUMULATIVE PASSENGER FLOW = ARRIVALS S ° ts Wig) eS TIME Standing and Seated Passengers Pigure 42 CUMULATIVE PASSENGER FLOWS > © VOLUNTARY STANDING TIME DEPARTURES ts TE, ‘components of Total Waiting Tine for Standing and Seated Passengers Pigure 5 If curve (1) is raised in the region, 1 < tit means that all those standing want to stand. If there are some who stand because there isnot sea, they will sit when a seat becomes availa and 50 all standees will be voluntary stanees.seat in series. te hatched region, Ry, represents the total waiting tine to standees. Curve 1 is also not representative of the true toarding behavior of passengers. All the seats are unlikely to remain cccupied shortly after the beginning of boarding at t,. curve 2 is « better representation of the situation at t > f, when not all avaliable seats my bbe occupied a2 sone passengers (impatiently) queue-up to board. For example, at tine £1, , passengers may be seated and h, passengers may ‘stand, even though a,<6. Thus the region 2, defined by curves 1, 2 and the cumulative departure curve, also represents waiting tine while standing. Movever, it 4a noteworthy that ell the passengers in R, are standing voluntarily while seats are available consequently, the total (compulsory) wating time while passengers have no choice but to stand, is the true definition of the region Ry. 2.2 optinal muster of Seats 1 our objective ts to Linit the maximum possible coopslaory standing tine, ttl t0 sone reasonable anount (say 15 minutes). the mmber of seats, §, can be chosen using rig. 5. We simply vary until t,t) = 15 ainotes. ‘this Ae somewhat axbitrary. We define the optinal nuster of seats as that vhich minimizes the sun of ‘the cost of the Lounge (proportional to it's area), the cost of seats and ‘the penalty for compulsory standing tine (proportional to the area of esion min Fig. 5): yah + yh * ype (area Ry) o vy, = cost of the lounge per unit area per aircraft departure, Yq ~ cost of a sent per atzcratt departure, Yp = cost of the penalty for each unit tive of compulsory standing tine per passenger,u ‘CUMULATIVE PASSENGER FLOWS. = 300} 270} LOAD MAXIMUM STORAGE |—s= 240. Ry REGIONS. (COMPULSORY. STANDING TIME) © 20. 40 60 8 100 120 TMEIMINUTES) variation in Compulsory Standing Tine with § Figure 6A. = Lounge area for passenger storass, and a = one plus the fractional increase in lounge area to allow pesuenger cirovlation and airline activities. 2 we substitute for A in (3) from (1), the total cost is given by YS + aMgm, Ot yp area Ry) w wien Yet enh - te a) me tem ay,m, can be interpreted as the minimin cost of storige per passenger and ye as the extra cost of providing a seat to a passenger. One ean also deiete @ fron the above formlations and simply ad@ a constant area to (3) for passenger circulation and airline activities (wart). ‘the first and third terms of (4) are increasing and decreasing functions of 5 respectively, while the a value of 5 at which (4) ia mininized, The area of region R, can be cond terms is independant of 6. There te thus estinated with « digitizer or a planineter for various values of $ and any curve F(t). Consequently, the sum ycS + yps(area Ry) can be evalusted for various § until the valve of § thet minimizes the sun 1s found. Consider a typical arrival curve for a scheduled flight (Ashford and Weight’) with a oad of 300 pessengers, shown in Tig. 6. the boarding rate 4s 840 passengers per hour, The maximin storage of passengers is 270. the compulsory standing time (represented by the regions R) are shown for a range of seate 120 to 270, Using the parameter values given in Table 2 and ‘the technique described above, the optimal number of seats for yp + $0.25, 50.50 and 1.00 per hour per passenger are 205, 240 and 250 respectively. 2.2 Lounge with Security check Lounges at gate-areival type terminals and many lounges for intemational aa1s able 2 Parameter Values 240 Paulin and Horonjert 400 aaa? ‘ransport cat 1s ‘reansport Canada” ao ‘Transport Canada’ « La Transport canada? Yy, S/oetert/alroratt . Yp $/our/pascenger 0.25 < c/etceratt departure 0,01" . Yg S78 Based on an initial investment of $2360 per sater*, 12% interest rate, 25 year Life, $100/meter? maintenance cost per year and 2000 Gepartures per year. ‘Based on an initial cost of $150 per seat, 128 interest rate, 6 years Alfe and 2000" departures per year.departures have dedicated security checkpoints at the entranc (12.7) 1A simple expression can be obtained for the optimal value of 5, if the tevivais to the lounge are thenselves the output fron @ security checkpoint te the lounge entrance at which « queue has forned Defore all the seats are fiiies im the Lounge, and Lf the queve Lasts util boeréing starts, Ten the arrival rate inco the loge can be uniform Goring & period of tine tyth tig. Ty This is rot a very comon scenario, However, the soltion of this epecial ase may give us done ineight to the more general problen. He eet ad te avait, - tpawikerm = = a ee m/teteed| eS vtace the airerate, 15 = rate at which passengers enter tho lounge from the security ceneckpoint vhen P(t)» substituting from (Sb) in (4) and using the first order condition for a ‘niniman, the optinal value of S 1s found to be B52 Qs B= SPY rg) > o ‘the optinat lounge size is obtained by substituting for § in (1), from (6) ‘the munber of standees at tine tyy s{l - 5/6) (yg/tp)+ Which is also the naximn ruber of standees (at any one tine), is sensitive to the penalty pr EB penaity increases for example fron $0.25 to $0,50/ote/ PMssenger, the number of standees 4s halved. the 8, obviously Gecresses a= the extra cost of providing © seat/passenger, yay increases. Since the vaiue of y, decreases as the frequency of use of the lounge increases, 6, co yeti be bigger at busy atzports. 16CUMULATIVE PASSENGER FLOWS uv ARRIVALS AT LOUNGE SECURITY CHECK - \ L >» QUEUE AT SECURITY CHECK Arcivals/Departures for Lounge with Security check Pigure 7Based on the paraneter values in Table 2, the maximum nunber of standees is 78 and the optisal nunber of seate Ls Q-78 where Q is the maximum storage ‘of pastongers in the lounge. 1 ts nore restistic to anne that the securttysheck rate will be ctoren to enmure Ua the queue at the check disappears before boarding starts, 1m ens ease, the neviod given in eection 2.2 can be uted. Rowever, if tyr the eppronioata area of rejion Nia obtained by reducing b in the honcrator of (0) by {0/2) (09/0). The related 8, 5 obtained by aividing the second term of (6) by {2 = (1/2) 6/0)41+ 5° 0 (eth -s/N1L = G72) 67/1) a) For example, the number of standece in the above estinate vith s/> = 0.48 As increased by 148 to 88, 2.4 Lounge Area cateulation art has given an example where a lounge is designed for Q = 261 people Gneluding visitors). Based on an average area of 1.1 m? for sitting and standing persons, he obtained a net area of 309 m4 for standing and seating. Presumably, Hart assuned 508 of the people to be standing, 4, 140 seats and M41 standees. Keeping a ~ 1 Sn (4a), our esleulations using (4a) and (63) give 82 standees and hence 200 seats. Taking the values of 1.4 a per seated person and 0.8 a? per standing person (based on 1.1m? average for seated tnd atanding persons) sed by Hart, and assuning the penalty ¥, = $0.25, gives an area of 345 m4 for the lounge. Tf the value for Y is $1.00 rer ‘passenger per hour, the net lounge area is 301 n*. The value of 209 m* obtained by Hart for the net lounge area is also obtained from our neting SE the valve of yp is 50.1425 per hour per passenger. A lounge without dedicated security checkpoint will be larger.2,0 OPFINAL LOUNGE SIZE (Wx of Alroraft and FLight Types) . ‘Given the constraint of boarding the passengers as close as possible to the departure tine, the capacity of the Lounge in terms of passengers, 0, i= fixed by a particular aircraft type combined vith a flight type (e.9. sntercontinental charter and Boeing 747). Consequently, @ 1s independent fof the mix of atreraft and flight types that will use the Lounge except to the extent that the particular afzerafe/€light combination that governed the choice of Q is a part of the mix. ‘The optinal nunber of seats ie hovever dependent on the mix of flights since the third term yps (area Ry) of the objective function (4), representing the compulsory standing time of passengers, will change with the alzeraft/flight type. Since Q ie fix |, 48 tn sufficient to mininize the first and third teme of (4) generatizes for the mix of alzcraft/tlight types 18 + Yp UP area RL), o (area Ry), = compulsory standing time for aircraft/flight -ype i Py ~ probsbtlity of a departure of aixcraft/tiight eye 4. ‘The value of § that mininizes (7) can be estimated in a manner sinllar to the mininization of (4), Keeping in mind that the area R is zere for i 1 with a, £5. 3.1 Lounge wien Security check consider several aizeraft/fLight types i with distinct arrival curw Pj(2). Me allow the discharge rate from the security check, sj, snd the aircraft poarding rate, bj, to be functions of t. Tf each £ aatiefies the sssunptions made in section 2.3 and Lf 0 > § for all i, the objectivefunction (7) is minimized when the (optinal) value of § is given by Soe Eee alt ney OE EDO ° = 0/2) 6971/18, 0 = 801. ° 1 0, =p for all 4, Le, the boarding and security check rates are uniform for all alreraft/flight combination 8,23 - ave 0 go B= C/G) a0) nore, the mean storage of passengers, arte ay 1£ the value of §, calculated from (8) or (10) is greater than sone of the Oy, there will be ne compulsory standing for these { and the sacond tam of (7) is zero for those aircratt/flight types. In this case 5, is recalculated fron (8) after setting the torm 9,9, of the minimn 9, to zero to account for the zero. compulsory standing tine, The procedure is repeated until the S, ie less than all the remaining 9,. Squation (10) cannot be used Sn this process because {t's derivation from (8) is Aependent on the existence of compulsory standing tine, Lie. 9, > 8, for aul i. Consider the cases shown in Table 3. case 1 ease 2 £0 pass, oe % 1300 0.6 0.1 2 250 0.3 oa 2100 oa 06 Atreratt/ELight Wx for tounge Designa case 2 ‘Taking Yp = $1.00 per hour per passenger and the remaining paraneter values from Table 2, Bq. (8) provides us with a firat estimate for the optima number of seats, $,(1) = 245, Since 5,(2) > Qy, we set Pypy = 0. Then the second estinate of the optimal number of seats, $,(2) = 261. since §,(2) > Oy 0, we set P30, = 7405 * 0. Then the final estinate of the optimal umber of seats, 6, ~ 266. The original ‘the compulsory standing time tera (Sb) inoreases with (Q,-5)%, rogaraiess of 0,28, unless St 1s set to zero for S20. false minimum is obtained because the objective function begins to prematurely increase with s. ‘There is a larger fraction of departures with relatively low load, in tnis case. §,(1) = 145 (293) and hence P30, 18 set to zero. Then §,(2) = 214, Since $5(2) < Op, Qy, the optinal number of seats, 5,» 212, 4.0 LOUNGE SHARED BY SEVERAL GATES A single lounge with one afzcrage gate 1s only adequate for small airports hore departures are spaced sufficiently far apart. However, most airports need several, Af not many, gates and approxinate methods for estimating the gate requirenent are available (Horonjefs’, Sandara and Wirasinghe!®}, the question then aries whether there is an economy of scale in sharing a owge anong seversl gates and ££ 2o what the capacity of the lounge should be consider n departures with identical cumlative arrival curves, P(t), 2 and door-closing tines. ‘then, the maximm accumlation of passengers is nQ, where Q 1s the maximm accumlation per departure, and it 4s independent of whether there is one lounge or n lounges. However, if the n departures are sufficiently far apert, auch thet the arrivals for 8 Aight do not begin until hoarding for the previous flight ie veil under way, the maximun sccunvlation of passengers is Q. The actual storage of passengers for n departure Se therefore in the range @ to ng.2 consider a lounge shared by several gates. Flights depart at various times, But the maximn storages Q and the tine periods t, during which passengers arrive at the lounge are the sane for all flights, (T9580 ‘me maximum accunulation of passengers in the lounge shared by several gates will occur at the beginning of boarding for a flight chosen such chat ‘the nunber of Lights for which passengers arrive, during a tine period ty before the beginning of boarding for that flight, is at a maximum, ¥ (Fis. ‘The aircraft departure rate would have peaked during the pericd t, as shown in Pig. Sb. If the peak can be assused to be symetrical in ty, the maximus accumlation of passengers is © 19/2, for large W. But, for small Ne (comparable with 5), the mean number of passengers per flight, as costed at tine ty, can vary considerably from 0/2. Taking M to be the stze of random sample from a uniformly distributed population in the range 0 to Q, ve take 0/2 plus to standard deviations ar an estimate of the upper bound fof the mean nusber of waiting pastengers per flight at tine ty co) + orem? a2 ‘then, the Jtimated upper bound of the passenger accumulation and hence the necessary lounge capacity in terms of passenger spaces, 1s otovay + evar . as ‘me Q in (22) can be replaced by § if # aix of various types of alreraft/flights is involved. tthe passenger epaces required for N separate lounges is Qi. Consequently, the passenger spaces in a conbined lounge az a fraction of the cpaces needed for separate lounges is Sos vm! nay 2 ror example, a saving of 178 in passenger spaces results by combining ounges for X= 3, and the saving increases to 501 as N+ ©,a3 CUMULATIVE PASSENGER FLOWS MAXIMUM ACCUMUL ATIO} ‘OF PASSENGERS te TiMe ‘¢) CUMULATIVE ARRIVALS/DEPARTURES FOR SEVERAL FLIGHTS iN nm TINE, DILIGHT DEPARTURES. FLIGHT DEPARTURE RATE, N FLIGHTS: N rues "2 u )ARRIVAL/DEPARTURE OURING A PEAK Yesinge Shared by Several cates igure 8roger confusion during peak periods, problens with etgning ond operational aiffieulties such as overlapping announcenents, will increase significantly with the number of gates that share a lounge. Purther, nost terminal geonstries do not favour the conbining of gate lounges. Thus very aarge lounges are inpracticsl. Pier-satellite and remote satellite locations are however guites for Lounge sharing. ‘The optimal mumber of seats can be estinated as explained in section 2.2. ‘the area R, in (4) $2 eatineted for a typical day by adding up the areas A, for each peak (Pig. Sel. Further, ¥, and Yq (hence Yq) are defined to be per day instead of per departure ‘AL the passengers in non-pesk period flights are Likely to obtain seats in the combined lounge. In addition to the reduction in the total munber of required passenger spaces, this ie the ther major advantage in sharing 2 tounge among several gates. Even when lounges are designed separately for ach gate, and subject to the security check being done elsevhere, the number of avattable seats can be increased by merging the Lounges along the boundari ‘the deterministic queueing theory approtch to estinating the passenger spaces for a lounge for a given aizcraft/flight type has been extended to the case of a mix of aircraft/flight types. Given the penalty for standing eimating the nusber of sea:e ‘that minimizes the sun of the cost of the Lounge and the total ponalty for standing, has been proposed. per minute per passenser,Y,, a method for ‘the average value of Yp 1s probably comparable with $0.50, wowever, i= vould be worthwhile to survey passengers to obtain a better estinats here ie Little agreement in the Literature (ae show in Table 2) regarding ‘the lounge areas per standing and easted passenger. There values too snovld be established with more precision.2s Te de shown that the area for lounges ean be reduced up to 50% by conbining tthe lounges of several gates. A method for estinating the combined lounge ‘size has boon presented. ‘This research vas supported in part by the Natural Sciences and Engineering Research Council of Canada under grant no. 8 4711. (1) B.C. Paulin and R, Horonjest, "Sieing of Departure Lounges in Meport Buildings", ASCE Transportation Engineering Journal, 267-277 (aay 1968). (2) &, Hanzaei (Transport Canada), "Techniques of Airport Terainal Planning in CATA ~ An Airport ‘Capacity Utilization Medel", Secosd Canadien Seninar on Systens Theory for the Civil Engineer, University of Calgary, Calgary, Canada (Hay 1984). (2) ‘Transport Canada, Departure Assembly Areas (Hold Rooms) , Report ‘e=82-01-800, Aueports and Construction Services Directorate, ‘eanepert Canada, Ottawa, Hay 1977. (®) W, Hart, the Airport Passenger Terminal, Wiley-Interscience, 1985, (5) G.P. Novell, Appiications of Queueing Theory, 2nd Edition, chapsan and ani, 188 (6) Pedaral Aviation Aamintstration, Aviation Demand and Airport Facility Requirenent Forecasts for Mediua Air Transportation Hubs, Washington, DiC., January 1968, (8, Moconjefe, Planning and Design of Airports, MeGraw HA21, 1975. (8), Ashford, "Passengers in Alzport Terminals", Airports international, (arch 1976) . (9). M, Ashford and Pall, weight, Atrport Engineering, wiley-interscience, 1579. (20) 5, Bandara snd §.C. WLrasinghe, "Airport Gate Position Eetination Under Uncertainty", paper presented to session 152, 67th Annual, Meeting, ‘Teanaportation Research Board, Wash. D.C., January 1983, 26 (a2) Rc, Paullin, Passenger Flow At Departure Lounges, Graduate Report, Institute of Transportation and Traffic Engineering, University of california, Berkeley, 196.Figure Figure Figure Figure Figure Pigure Figure Figure List of Figures ‘typical Boarding Sequence [Paullin and Horonjeft!) Arrival and Departure of F: Rizerafe/Elight Types rengers for Two Different Departures From a Comon-Server Security Check Standing and Seated Passengers Components of Total Waiting Time for Standing and Seated Variation in Compulsory Standing Time vith § -erivale/Departures for Lounge with Security check Lounge Shared by Several cate