Airport Congestion Management in a Cost Efficient Way

Transcription

1 T E C H N O L O G I C A L E D U C A T I O N A L I N S T I T U T E O F K A V A L A Airport Congestion Management in a Cost Efficient Way by E. Giarmatzides March 2006 M.Sc. Finance & Financial Information Systems School of Finance, University of Greenwich

2 Airport Congestion Management in a Cost Efficient Way Terminal Passenger Processes by Elias Giarmatzides A dissertation submitted in partial fulfillment of the requirements of the degree of Master of Science University of Greenwich March 2006 Supervisor: Dr. Elias Sarafis 2

3 ABSTRACT Airport passenger traffic has increased and will continue to do so in the next coming years. This enormous increase in air passenger movements, plus some uncontrolled events such as bad weather or seasonal air traffic variations, overloads the air transportation system and generates congestion and costly delay problems worldwide. Airlines, passengers and airports alike incur these resulting costs. In an attempt to manage airport congestion and associated costs, international air transportation organizations, airline and airport administrators mostly concentrated on the analysis of the effects of airspace and runway parameters of the air transportation system in creating the problem. Not much emphasis was given to analyse the contribution of inefficient airport-terminal passenger facility processes into the effort to solve the problem. The purpose of this dissertation is to review literature for existing methods that enhance airport terminal capacity but mainly to propose alternative ways that could result to more efficient airport-terminal passenger flow processes and thus manage airport congestion problems in a cost efficient way. A Computer based method will be implemented, as computer and information technology systems are capable in manipulating fast vast amount of data. A computer based software simulation model of terminal passenger flow processes will be proposed that will assist us to analyze and decide on efficient operation of passenger terminal facilities. Witness software simulation package will be used to animate the flow of passengers inside the terminal building and portray graphically the layout of the terminal facilities. A number of different simulation scenarios will be executed to enable us to determine the most critical parameters of terminal facilities that primarily affect the value of the process time. Process time denotes the value of time required for the completion of the entire passenger flow procedure. The model will be verified and validated by data obtained from air transportation organizations such as ICAO, IATA, Eurocontrol, FAA and appropriate airline standards if publicly available. The output results obtained 3

4 by the model will be fetched into excel spreadsheets for cost evaluation based on economic analysis of waiting lines. Cost evaluation will focus on key cost components that primarily affect the total cost of servicing air passengers inside the terminal building such as passenger value of time cost, airline service cost and delay cost. It will be shown that by efficiently utilizing existing passenger terminal facilities the demand for continuous growth in air passenger traffic can be accommodated in a cost efficient way while maintaining the proper level of standards. 4

5 ACKNOWLEDGEMENTS I would like to thank the people who spent time discussing relevant subjects about my dissertation and gave me important inputs on how to proceed. Dr. Sarafis Elias, professor of the IT stream for accepting to be my dissertation supervisor and giving me the motivation to go on. Dr. Theriou Nikolaos, the supervisor of the postgraduate program for taking time to discuss economic and financial aspects that can be implemented into the dissertation. Dr. Chatzoglou Prodromos, professor of the postgraduate program, for taking time to make helpful suggestions on how to proceed on the chosen topic. Mr. Nikoleris Nikolaos, mathematician for his support on the model implementation from the programmers perspective. 5

6 TABLE OF CONTENTS LIST OF TABLES viii LIST OF FIGURES ix LIST OF ABBREVIATIONS xi CHAPTER 1 Introduction Air Traffic Data Delays and Congestion Delay Codes Delay Causes Capacity-Congestion Airport Congestion Causes Congestion-Delay Costs Methods to Reduce Congestion-Delays 9 CHAPTER 2 Terminal Facilities and Passenger Processes Terminal Facilities Level of Service (LOS) Passenger Processes Ticket Process E-Ticket Process Internet Ticket Process Baggage Control Baggage Screening Positive Passenger Baggage Match Post September 11 Process Requirements Check-in Process Check-in Kiosks Security Checks Customs and Immigration Boarding Simplifying Passenger Travel Program 27 CHAPTER 3 Research Methodology and Data 30 6

7 3.1 Simulation Model Methodology Simulation Model Software Model Assumptions Model Passenger Flow Description Input Data Model Parameters Model Implementation Economic Data on Cost Values Cost Computation Formulas 40 CHAPTER 4 Results and Evaluation Simulation Results Without Ticket-Baggage Passenger Scenarios Main Simulation Model Results Model Results Analysis Economic Results Analysis Model Implementation and Study Limitations 53 CHAPTER 5 Conclusions Recommendations for Future Research 55 APPENDIX A: Standard IATA Delay Codes (Annex 3) 56 APPENDIX B: Witness Model Activity Layout 57 APPENDIX C: Witness Demo Outputs 61 APPENDIX D: Main Simulation and Cost Results 71 APPENDIX E: Witness Software Outputs Statistics (140 Pass.1 Check-in) 89 BIBLIOGRAPHY 93 7

8 LIST OF TABLES Table 2.1: LOS Standards 16 Table 2.2: Space Guidelines for Capacity 17 Table 3.1: Aircraft Type Passenger Capacity 36 Table 3.2: Data on Airport Terminal Facilities 37 Table 3.3: Unit cost values 42 Table 4.1: Demonstration Simulation Scenario 45 Table 4.2: Without Ticket-Baggage Scenario 46 Table 4.3: Cost efficiency values-optimal Check-in number (Domestic) 52 Table 4.4: Cost efficiency values-optimal Check-in number (Intern.) 52 Table C.1: Entity Statistics Report 61 Table C.2: Queue Statistics Report 61 Table C.3: Activity Statistics Report 63 Table C.4: Conveyor Statistics Report 66 Table D.1: Passenger number per aircraft type 71 Table D.2: Domestic Flight Type Scenario Cases 71 Table D.3: Domestic Flight Type Cost Calculations 74 Table D.4: International Flight Type Scenario Cases 77 Table D.5: International Flight Type Cost Calculations 80 8

9 LIST OF FIGURES Figure 1.1: Air Traffic in Figure 1.2: European yearly traffic variations 3 Figure 1.3: Total ATFM Delay 5 Figure 1.4: Primary Departure Delay Causes (Source CODA, 2004) 6 Figure 1.5: Traffic variations 8 Figure 2.1: Ground resources for passenger handling 12 Figure 2.2: Arrival and service patterns at airport facilities 13 Figure 2.3: Terminal passenger flows 18 Figure 3.1 Simulation process 32 Figure 3.2: Model facility counters and passenger flows 35 Figure 3.3: Waiting line cost curves 41 Figure 4.1: Terminal passenger flow 44 Figure B.1: Model Title 57 Figure B.2: Queue Process 57 Figure B.3: Ticket Counter 58 Figure B.4: Baggage Control Counter 58 Figure B.5: Check-in Counter 59 Figure B.6: X ray Counter 59 Figure B.7: Passport Counter 60 Figure C.1: Queue Statistics Chart 62 Figure C.2: Activity Statistics Chart 64 Figure C.3: Activity Statistics Chart Status 65 Figure C.4: Conveyor Statistics Chart 67 Figure C.5: Conveyor Statistics Chart Status 68 Figure C.6: Check-in Utilization (%) 69 Figure C.7: Total Time in Queues per Facility (%) 69 Figure C.8: Total Time Spend per Passenger State (%) 70 Figure C.9: Service Time per Facility (%) 70 Figure D.1: Number of Passengers Figure D.2: Number of Passengers Figure D.3: Number of Passengers

10 Figure D.4: Number of Passengers Figure D.5: Number of Passengers Figure D.6: Number of Passengers Figure D.7: Number of Passengers Figure D.8: Number of Passengers Figure D.9: Number of Passengers Figure D.10: Number of Passengers Figure D.11: Number of Passengers Figure D.12: Number of Passengers Figure E.1: Passenger Flow 89 Figure E.2: Baggage Control Statistics 90 Figure E.3: Check-in utilization (%) 90 Figure E.4: Total Time in Queues per Facility (%) 91 Figure E.5: Total Time Spend per Passenger State (%) 91 Figure E.6: Service Time per Facility (%) 92 Figure E.7: Total Time Spend per Passenger State (%) (No. of Check-in 3) 92 10

11 LIST OF ABBREVIATIONS ACI Airports Council International AEA Association of European Airlines ATAG Air Transport Action Group ATFM Air Traffic Flow Management AUC Air Transport Users Council BAA British Airport Authority CFMU Central Flow Management Unit CODA Central Office for Delay Analysis CUSS Common Use Self-Service CUTE Common-Use Terminal Equipment FAA Federal Aviation Administration IATA International Air Transport Association ICAO International Civil Aviation Organisation ITA Institut du Transport Aérien PPBM Positive Passenger Baggage Match RFID Radio Frequency Identification 11

12 CHAPTER 1 Introduction In today s business world the demand for air transportation has started to increase again to accommodate the needs for increasing volume of air passenger movements and goods. Air industry seems to recover after the setback has suffered due to the events of September 11 th and more recently War in Iraq and the outbreaks of the SARS virus. Recovery started to appear in 2003 and continued in 2004, which was a good year in terms of passenger figures and cargo movements. In an address to the 29 th Annual Aviation Forcast Conference of the Federal Aviation Administration (FAA), James May, president and chief executive of the International Air Transportation Association (IATA), said (Jar, 2004:9): As we enter 2004, we leave one of the most cataclysmic periods that the industry has ever faced; 11 September, SARS, the downturn in the US economy, the Iraq War are just a few of the events that have shaped our recent history. Airports Council International s (ACI) statistical indicator for 2004 shows double-digit growth rates for passenger traffic and 10% increase for cargo. According to the latest FAA forecasts, the commercial airline industry will grow at an annual rate of 4.3% between now and 2015, with regional airline passenger traffic outpacing that of the major long-haul air carriers (Jar, 2004). Boeing (2000) and Airbus (2000) forcast that, the world's fleet will double by the year 2019 in order to accommodate air traffic growth rates. New market trends in worldwide economy, e-commerce, institutional reforms, and airline deregulation, airport privatization, trade expansion and globalization have caused the boom in air travel and services. As the cost of air travel, due competition between airlines continues to decrease, the number of people flying worldwide for business or pleasure will continue to rise. Economic growth and prosperity, passengers who demand a more convenient, safer and faster way of transportation have also increased air travel volume. One of the leading airport planners, Horonjeff (1994) says: People have been brought closer together and so have reached a better 12

13 understanding of interregional problems. The industry has found new ways of business. The opportunity for more frequent exchange of information has been facilitated and air transport is enabling more people to enjoy the cultures and traditions of distant lands. Air transportation is the fastest growing sector of transportation in developed countries. This increasing demand in air transportation puts significant strain on airport and airspace resources as airport infrastructure and airspace capacity are not expanding at a similar rate. It is this growth in demand, further than the system capabilities and resources that causes the problem of congestion and generates delays at major airport terminals worldwide. The Air Transport Action Group (ATAG), which is funded mainly by Airbus Industries, Boeing Commercial Airplane Group and IATA, has acknowledged that airport infrastructure constraints generate major delays, which consequently have economic cost that must be taken into account. Airport terminals, which play a key role in the commercial aviation system some times fail to accommodate this growing demand in air transportation due to lack of capacity or inefficient operation. It is a challenging task for airport planners, developers and operators to keep pace with the rapidly growing demand for air transportation (Dempsey, 2000). 1.1 Air Traffic Data Since the 1960's, the aviation industry has experienced rapid growth worldwide and even greater after deregulation came into effect in the 1970 s. According to ACI (2004) Annual Report the world s airport traffic data as shown in Figure 1.1, were 3,656 millions passenger, 73,867 thousands cargo and 62,746 thousands aircraft movements in total. Figure 1.1: Air Traffic in 2004 (Source ACI, 2004) 13

14 In Europe based in Eurocontrol s data traffic in 2003 grew +2.8% on average, 0.5% above the 2000 traffic level as shown in Figure 1.2 (CODA, 2004). Figure 1.2: European yearly traffic variations (Source CODA, 2004) It is clear that despite the recent negative events -depressed global economy, terrorist attack of 11 September 2001, Iraqi War and SARS- air traffic will continue to grow. According to the ACI World Forecasts, global airport passenger traffic will grow by over 4% annually up to 2020 to some 7 billion passengers. This growth is the highest compared to any other means of transportation. Other organizations, such as International Civil Aviation Organization (ICAO) and IATA, are even more optimistic indicating that European air traffic is now expected to increase at the rate of 5% to 6% per annum, doubling by Air transportation is the fastest growing sector of transportation and one of the fastest growing sectors of the world economy. As a consequence of this growth in air passenger traffic airport terminal operational capacity is put under considerable strain resulting to congestion and associated costly delays. 14

15 1.2 Delays and Congestion As the demand for air travel increases, congestion and delays in the air traffic system become more and more usual. Delays may have as a consequence congestion at airports, but the relationship may not be entirely in that order as also congestion could result in delays. According to international standards delays are only recorded when an aircraft exceeds its scheduled arrival or departure time by at least 15 minutes. Delays or increased congestion at busy airports results in significant financial and environmental inefficiencies Delay Codes Air transportation agencies in an effort to standardize internationally data kept on air transportation delay causes agreed to implement a common delay coding system. The delay coding system that was initially implemented by IATA is the most widely adopted coding system in order to record delays. Also some airlines developed supporting delay coding systems to complement the IATA coding system according to their specific needs and operations. In the IATA coding system, published in Annex 3 as can be seen in Appendix A, delay causes are categorized into 100 codes (No. 00- No. 99) and grouped into 12 major groups (CODA, 2004): 1. Others 2. Passenger and Baggage 3. Cargo and Mail 4. Aircraft and Ramp Handling 5. Technical and Aircraft Equipment 6. Damage to Aircraft & EDP/Automated Equipment Failure 7. Flight Operations and Crewing 8. Weather 9. Air Traffic Flow Management Restrictions 10. Airport and Governmental Authorities 11. Reactionary 12. Miscellaneous 15

16 1.2.2 Delay Causes The main source of data for delay causes come from agencies such as IATA, ACI, Association of European Airlines (AEA), and the Centre for Delay Analysis (CODA) at Eurocontrol and their results are published annually. Even that these data are publicly available it is considered highly sensitive and confidential to reveal detail information concerning airlines or airports as units. The CODA at Eurocontrol started publishing monthly and annually delay reports in 1996, by collecting and comparing data from different sources such as IATA, ACI, AEA and air traffic flow management data reported by the Central Flow Management Units (CFMU) within Eurocontrol. According to Eurocontrol for the year 2000, 77% of the delays were airborne delays and 23% ground delays. In 2003, the ratio changed significantly with 54% of the delays attributed to en-route phase of the flight compared to 46% attributed to airport delays. Ground delays were doubled. For 2004, analysis of the delay causes shows that en-route accounted for 51% of the delay while ground delays for 49% (CODA, 2004). In a more detailed consideration Weather accounted for 20%, ATC Equipment 5.5%, ATC Staffing 5% and Airport Facilities 15% of the delay causes as shown in Figure 1.3. Figure 1.3: Total ATFM Delay (Source CODA, 2004) 16

17 An analysis for 2004, of the primary departure delay causes indicates that 51% of delay causes were due to airline operational activities 4 points more than 2003 (CODA, 2004). Airports accounted for the 19% of the primary departure delay causes 2 points less than In total airlines and airports accounted for the 70% of primary departure delay causes in 2004 two points more compared to 2003 data as shown in Figure 1.4. The graph indicates that the majority of delays were related to airport and airline difficulties. Figure 1.4: Primary Departure Delay Causes (Source CODA, 2004) 17

18 1.2.3 Capacity-Congestion Knowledge of airport s capacity enables airport and airline managers to handle efficiently forecasted aircraft movements and passenger flows, and to estimate the size of delays in the system as demand varies (Ashford and Wright, 1992). Some of the important aspects that define an airport s capacity are the number of terminal facilities, the number of gates, size of aprons and how efficiently they are utilised (Reynolds and Button, 1999). Congestion occurs when demand approaches or exceeds capacity. Lack of airport capacity has been forecasted by the FAA to be one of the most serious constraints to the growth of commercial and private aviation (Wells, 2000). The mismatch between the required and existing capacity on the air or ground, especially during peak traffic periods, is what creates congestion problems at airports. Insufficient capacity during peak traffic periods imposes significant operational, economic and environmental constrains to providers and users of air transportation network alike Airport Congestion Causes Congestion problems at airport terminals occur for several reasons and can be attributed to the airports themselves, the airline schedules, travel agents and passengers. Traffic variability, weather conditions and aircraft characteristics also can be sources of congestion. When airport s managers make seasonal plans on fight schedules and slot allocation to airlines, in order to be competitive to other airports, overlook sometimes that they overload airport s terminal facilities. Large airports that operate as hubs could suffer congestion problems as their capacity deteriorates when a lot of planes must be serviced in short periods. Large number of local and transfer passengers must be processed efficiently through the terminal. Airlines in general do not take into consideration air traffic congestion when plan their flight schedules. As business enterprises, they must schedule antagonistically according to their operational strategy in order to compete financially and satisfy their customers. They design their flight schedules 18

19 without considering the costs that they could suffer or impose on other airlines. A minor delay on a strict flight timetable schedule would generate costly reactionary delays as the accumulation of delayed passengers overloads the capacity of the terminal facilities. Travel agents and large tour operators are also components of the air transportation industry. Travel agent s behaviour as they handle passengers in large groups, especially during the busy summer period, affects a great deal airport s terminal capacity and could cause congestion. Another reason that airport terminals could experience congestion is that more and more business passengers travel at times that are considered convenient to them. Business people prefer to fly early in the morning and late in the afternoon. The same effects have also the larger numbers of passengers that travel during the summer period for vacation purposes. This traffic variability -patterns of asymmetrical traffic- as shown in Figure 1.5 could lead to congestion during these peak flight periods if terminal resources are not managed efficiently. Weather conditions are also important as they could add problems to airport s capacity mainly during winter periods. Low visibility, high winds, and the accumulation of snow or ice reduce the capacity of the airport runways. Weather created operational problems usually sustain for longer periods thus affecting the capacity of the terminal area and so creating congestions. Figure 1.5: Traffic variations (Source EUROCONTROL, 2004B) Aircraft industry tendencies in developing huge aircrafts, such as the A380 developed by Airbus, will also have an impact on airport capacity. While these huge aircrafts could relieve airside congestion by carrying more passengers 19

20 with fewer flights, they could also put significant operational constrains on current terminal infrastructures that could only be relieved with expansion and efficient management. To be able to accommodate up to 800 passengers per flight, ticket processing, baggage handling and check-in processing facilities will need to increase or operate efficiently in order to avoid congestion. 1.3 Congestion-Delay Costs Delays due airside or landside congestion are very costly, both to air travel providers and customers. Delay congestion problems are translated into increased operating costs for airlines and valuable time losses for passengers. The total air congestion costs were estimated around 5.5 billion for Europe in 1995 and $4,5 billion for U.S.A in 1998 (ATAG, 2000). In Europe for the year 2003 costs incurred by airspace users due to ATFM and associated reactionary delays are estimated at 400 million for en-route delays and at 400 million for airport ATFM delays (Eurocontrol, 2004B). It is estimated that a top-10 European carrier suffers $100 to $400 million of delay costs annually which significantly degrades the profitability of airline business as well as its business competitiveness (Suzuki, 2000). At Madrid airport in 2000 the sum of congestion costs endured by airlines and passengers were estimated at million, 70% due to airlines extra costs and 30% due to passengers lost time (Nombela et al. 2004). Even though different approaches and assumptions are employed by organizations in estimating congestion and delay costs, the above estimated numbers point out the size of the problem of flight delays. 1.4 Methods to Reduce Congestion-Delays Air transportation industry agencies and economists search for efficient ways on how to alleviate the problem of airport congestion thus avoiding delays and associated costs. Proposed solutions to the problem of airport congestion and the associated delay costs vary according to the planning horizon (Panayiotou and Cassandras, 1999). Long-term solutions involve building new airports, which 20

21 would allow passenger traffic to be spread among more airports than to have a few airports extremely congested. This is considered the most expensive and time-consuming option. Huge capital investments would be required to build the entire facility but in the long run economic growth will be created for the surrounding community. Medium-term approaches would consider expansion of existing airport facilities that could also require large capital expenditures and time. Another approach is to invoke airport congestion pricing for periods where peak-loads of demand are observed (Daniel and Pahwa, 2000). Finally, short-term solutions are based on utilizing existing airport terminal facilities efficiently. The European Commission's White Paper on European Transport policy for 2010 clearly states that: In response to the growth in traffic, it is time to rethink how airports operate in order to make optimum use of existing capacity. Airport and airline operators must search for methods and procedures that optimise the use of existing terminal passenger facilities in an effort to expedite passenger processes. This can be accomplished through the use of operations management methods and information s systems technology. Studies concerning airport passenger models (Tosic, 1992) and estimation of capacity for passenger terminal facilities (Brunetta et al. 1999) can assist operations managers on that attempt. Methods and practices that focus on improving infrastructure utilization by making more efficient the passenger flow inside the airport terminal building is a subject of research that can be found in relevant literature (Setti and Hutchinson, 1994). Every procedure that a passenger has to go through in order to board the aircraft is thoroughly examined giving particular attention to time consuming processes e.g. the check-in process (Chung and Sodeinde, 2000). A number of methods that assist airport operators and managers to optimise airport terminal passenger processes in a cost efficient way include computer based network queuing models, and simulation models. Simulation models are especially useful when analyses of the operations of passenger terminal activities are to be performed quickly at a relatively detailed level with the use of few available data (Jim and Chang, 1998). The objective of this research is to manage airport congestion problems in a cost efficient way by operating effectively existing terminal passenger 21

22 processing facilities. In that attempt a computer based simulation model of airport terminal facilities and related passenger processes will be implemented. The economic benefits will be considerable, both for the users and providers of air transportation. The contents of each subsequent chapter of the research are listed below. Chapter 2 presents a detailed review on the issues of the airport terminal facilities and the standard practices that are employed for passenger processing. Also at the end of the chapter the initiative for the SPT program is illustrated. Chapter 3 describes the methodology that will be followed to determine the optimal configuration of passenger facilities for the proposed terminal model for cost efficient passenger processing. It presents the simulation of the flow of passengers inside the model terminal building implemented with the use of Witness software package plus the computation of the costs involved, based on economics theory of waiting lines (queues). Also provides descriptions of the parameters and the data and data sources referenced that will be used to run our model and calculate the relevant costs. Chapter 4 presents the analysis of the simulation model outputs, resulted from the various terminal facility configuration scenarios performed. Also the evaluation of the different costs calculated that would assist us to determine optimal configuration of facilities for cost efficient passenger processing. Chapter 5 outlines issues covered in the previous chapters and presents conclusions and suggestions for further research. 22

23 CHAPTER 2 Terminal Facilities and Passenger Processes Airports are one of the major components of the air transportation network system the other being the air traffic system. Some of the functions airports provide are access for taking-off, landing and maneuvers of aircraft and also ground resources for passengers. The typical ground resources that handle passengers from the time they arrive and until they depart are shown in Figure 2.1 (Horonjeff, 1994). Figure 2.1: Ground resources for passenger handling (Source Horonjeff, 1994) One of the main components of airport ground resources along with access interface and flight interface areas is the passenger terminal building. In fact, one of the key functions of the passenger terminal is the change of movement type (Ashford and Wright, 1992). That is the accumulation of passengers who come to the airport in small groups to form batches, which will be carried together in an airplane and split into small groups again at the destiny airport. The passenger terminal provides the appropriate facility infrastructure for this 23

24 change of movement type for passengers to take place. Terminal facilities are expected to process passengers efficiently and securely always, and also during peak hours of demand. In support to that effort, of more efficient passenger processing through the terminal facilities, an initiative is underway by a group of organizations called the Simplifying Passenger Travel Program (SPT). 2.1 Terminal Facilities The terminal passenger facilities are divided into processing facilities (ticketing, baggage control, check-in, security screening and passport control), holding facilities and flow facilities. Terminal facilities, that enable all time efficient passenger processing in terms of process time is not always the case especially during peak periods of demand. The capacity of the terminal facilities is usually set such that a certain level of service will not be achieved during a few hours of the year (Ashford and Wright, 1992). For any terminal facility the passenger arrival process associated with a departure flight can be described by a curve as shown in Figure 2.2. Figure 2.2: Arrival and service patterns at airport facilities According to queuing theory the number of servers in a system is the primary design variable (Render, 2000). Everything else being equal, the more servers available the less waiting time for the customers. Likewise in an 24

25 airport terminal system where different facility counters represent servers, it is of grade importance to determine the optimal number of facility counters required. This way we will be able to minimize total cost, by keeping a balance between passenger and airline costs, and still maintain efficiency on passenger processing. Other important variables that can also influence the performance of the terminal facilities are: Configuration design (layout) of the facility counters Way they are operated Number of queues Since from all the terminal facilities, check-in and baggage drop facilities are the most critical for efficient passenger processing (ICAO, 1999; BAA, 2001; Lemer, 1992), the above three performance variables will be illustrated for this type of facilities only. In general, three configuration styles in the design of check-in and baggage drop terminal facilities are utilized (Horonjeff and McKelvey, 1994): linear, pass-through, and island. The linear type is the most frequently used check-in counter configuration, as its multi-purpose positions reduce the number of serving stops for passengers and afford flexibility in staffing. The pass-through type, which can reduce passenger cross-circulation at check-in areas and provide increased baggage transfer capabilities. The island type that can provide interchangeability between multipurpose or specialized functions as it is configured in the shape of a U around a single conveyor belt. This type of check-in configuration is mostly encountered at large-scale airports. They are operated according to one of the two following methods or a combination of both: the common-use method and the dedicated method. The type of method employed each time depends on the requirements and the nature of the offered services. A check-in facility operating with many common-use counters will allow relatively higher utilization of the counters compared to an operation with dedicated counters. Thus, the most efficient check-in configuration and operation is a long continuum of counters at any of which passengers may check-in or drop baggages irrespective of their airline 25

26 or flight number. The largest airports use a common-use check-in operation system in order to enhance utilization and efficiency. As far as queue performance, using a single queue for several check-in counters instead of one queue for each check-in counter is usually more efficient because it is perceived by passengers that waiting time is equitably distributed among them. Moreover, in cases where passenger-processing needs are homogeneous a single waiting line can provide shorter and less costly average waits. Additionally, significant factors that will determine the optimal number of check-in counters and space requirements for cost efficient passenger processing are: Rate and distribution of passenger arrival at the check-in counters Type of aircraft for handling (Seat capacity) Type of flight (Domestic, International, Charter) Level of service standard Information Systems (Computers for passenger reservations) Equipment reliability and performance (Baggage belt-conveyor) It is important to realize that the actual required number of counters to serve the demand during peak departure times is greater than the forecasted number by airport or airline managers. Batches of passengers arrive to the check-in counters during short periods of time as close to the departure time as they can, so check-in counters may not be operated to full efficiency during a peak. This will be the case even with a common-use operation method. Thus, each airline must decide how many check-in counters to operate and how many service agents to engage especially during peak operating hours. Passenger arrival distribution at the airport terminal prior to flight time and aircraft passenger capacity under service are primary factors that must be considered when assigning the number of check-in counters, together with check-in service rates Level of Service (LOS) Passenger terminals and passenger processing facilities must comply with airport operation regulations and standards set by international agencies. 26

27 According to ICAO (1999) recommendations a passenger terminal to process travellers efficiently must follow the following standards: 60 minutes for international departure passenger processing from presentation at first processing point to the scheduled time of flight departure 45 minutes for international arrival passenger processing from disembarkation to completion of last clearance process Another organization IATA has also set some standards regarding the quality and conditions of service of one or more terminal facilities for the processing of passengers. These standards refer mainly to typical measures of service level in terms of waiting time, processing time, walking time, and space availability for each passenger. For airports maintaining higher level of services implies higher costs on one hand but higher charges to airlines that make use of these facilities on the other. To indicate the level of services (LOS), a set of letters from A (excellent) to F (unacceptable), are used in terms of flow, delays and level of comfort as shown in Table 2.1. Typically, level C is recommended as a minimum and level D is considered tolerable for break down periods. Levels of service standards in terms of capacity for processing facilities according to IATA are shown in Table 2.2. Table 2.1: LOS standards (Source IATA, 1998) LOS Level Description A Excellent Free flow, no delays, excellent comfort level B High Stable flow, very few delays, high comfort level C Good Stable flow, acceptable delays, good comfort level D Adequate Unstable flow, passable delays, adequate comfort level E Inadequate Unstable flow, unacceptable delays, inadequate comfort level F Unacceptable Cross-flow, system breakdown, unacceptable comfort level Airports and airlines have also developed their own performance standards criteria for efficient passenger and baggage processing such as: Congestion space standards Queuing Times Processing times Transfer connection times 27

28 Walking distances Information Systems Equipment reliability/performance Table 2.2: Space Guidelines for Capacity (Source IATA, 1998) LEVEL OF SERVICES STANDARDS IN SQUARE METER PER OCCUPANT A B C D E F Check in System Wait/circulate break down Hold room Baggage claim area Pre-inspection Passenger Processes The sequence of processing activities that a passenger has to go through in order to board a plane mainly takes place inside the terminal building and is considered as the terminal s main function. With the exception of ticket buying activities, which mostly take place at airline or travel agents offices located outside airports, and a very few off-airport baggage check-in activities, every pre- or post flight activity related to the air trip is carried out at the airport terminal building. Flows of travellers that arrive in an airport terminal are divided into two main streams as shown in Figure 2.3. Arriving passengers and departing passengers. Arriving domestic passengers will reclaim their baggage and will leave the airport while international passengers furthermore have to proceed through customs and immigration. Departing passengers on the other hand have to go through more time consuming processes that greatly affect terminal capacity and efficiency and could be the cause of delays. Departing passengers with their tickets on hand process their baggages through a baggage control security system to ensure that not potentially dangerous objects are carried. In some airports, usually large ones, this 28

29 activity takes place after the check-in procedure in an area not accessible to the passengers. Then passengers, with the appropriate verification documents and tickets, proceed to the check-in counter where an airline employee checks their tickets, acknowledges their enlistment in the boarding list through an information system, issues boarding passes, and routes baggage s with the appropriate tags, to be loaded on to aircraft. Then they go through a security search process for themselves and their carry-on baggage and enter into the departure hall, which is a secure waiting area. In addition for international travellers their passports or some other ids are checked. Finally, at the boarding gate passengers present boarding passes and ids to an airline employee and board or carried to the plane. Airplane access area Transferees Arriving passengers Departure hall DOM Passport control DOM INT Passport control Security check INT Check-in Baggage claim DOM INT Baggage control Customs Immigration Ticket counter Departing passengers Terminal access area Figure 2.3: Terminal passenger flows 29

30 Next the departing passenger processes will be reviewed in depth, to comprehend all the aspect of the different processes, as they significantly affect time and cost efficient operation of airport terminals Ticket Process The air traveller ticket besides being a receipt for the holder gives passenger the right by obtaining a boarding pass at the terminal check-in to travel to a pre specified destination on a certain flight. The passenger hands the ticket, at the check-in counter, to an airline officer to get a boarding card and also verifies with the use of some id that is the same person written on the ticket. Passenger ticket resembles a contract between two parties, passenger and airline, with certain obligations and rights for each party involved. Each ticket is issued to a specific person, is not transferable, and the holder could get a refund in a case of flight cancellation. The ticket purchasing process has been improved greatly through the use of Information Technology thus providing benefits to all parties involved. Airlines and travel agents save the costs of sales staff. Airports save terminal physical space as the requirements by airlines for ticket counters are reduced. Nowadays ticket purchasing is mostly considered as a travel process that takes place outside the airport terminal building. An airline passenger can arrange to purchase and pay for a ticket from home or office through a travel agent or directly with the airline. IT enables travellers to make travel reservations by phone, through Internet or E-Ticket Process Prior to IT and computer communication s evolution, the paper air ticket was a basic tool for passenger processing. In today s high-tech environment this necessity it tends to become almost obsolete as passenger s needed id information concerning flight reservations could be entered into the computer and verified in a few seconds through a reservations database. All major world airlines are eager to fully implement the electronic ticket (e-ticket) procedure despite existing limitations, as it enables them and their passengers to reduce cost and save time in booking and check-in. United 30

31 Airlines (2000) reported in May 2000, that more than 60% of the tickets used by its customers were e-ticket. Some limitations that still exist have to do with the right claimed by a number of airlines to cancel issuance of advanced boarding passes due to security reasons (ICAO, 1999). Another drawback is that e-tickets are generally limited to the airlines that issue them, as there is no cooperation between them. If an air trip involves flying on more than one airline then conventional paper tickets must be used. To overcome these limitation airlines must cooperate in order to integrate their ticket sales systems following the model of Air Canada and United Airlines as they already offer the convenience of e-tickets to their passengers if they fly on both airlines during a trip (United Airlines, 2000). Another cutting edge initiative employed by some airlines and based on the e- ticket concept is the utilization of smart cards (ICAO, 1999). All the essential information concerning the flight along with some passenger personal data relating to security reasons can be stored electronically in the smart card Internet Ticket Process Travellers that have access into the Internet can make reservations and purchase their tickets online from airlines or travel agents with sites in the web. Internet can save passengers, travel agents, airlines and airports a significant amount of money and process time. Online ticket sales will continue to increase as Internet access enables the traveller to choose between many alternatives offered by different service providers at a price competing environment. In 1998 Northwest Airlines reported that had increased online ticket sales to 800 a day compared to 36 only 18 months earlier and Delta Airlines expected to increase considerably online ticket sales well above 2% of the total sales (Air Transport World, 1998) Baggage Control In spite of the September 11, 2001 or earlier catastrophic events air transportation is generally considered as the safest form of transportation but still remains one of the most vulnerable despite the extra strict safety procedures. As a result of this inherited vulnerability, that affects not only the 31

32 life of hundred of passengers but also the smooth operation of airports and airlines, flying aircrafts have become the favourite target of terrorist actions. To overcome this obstacle globally, standardized baggage control security procedures have to be followed to ensure that no passenger baggage loaded to an aircraft contains any kind of hazardous objects that could endanger the flight. The two most common processes set by international air transportation agencies with that task are: baggage screening and baggage reconciliation also known as Positive Passenger Baggage Match (PPBM). Furthermore, after September 11 baggage-screening requirements became even more rigorous Baggage Screening Baggage screening devices and processes are employed at airport terminal facilities to make sure that no dangerous objects that could jeopardize any flight are transferred into airplanes in passenger s baggages. Integrating them into the sequence of the air passenger processes in a way that could not add extra time to the whole procedure is a big challenge for airport and airline managers. In confirmation of that the British Airport Authority (BAA) some years ago conducted a pilot project at Glasgow Airport with an X-ray system capable of processing up to 1,200 bags per hour per line (Aldo, 1993). Experiences from various incidents have shown that X-ray devices alone cannot detect all the different kinds of dangerous material like explosives. Additional more innovative detection devices, such as the ones that use cranium axial topography (CAT) technology are needed (Marsh, 1997). New screening machines combined with artificial intelligence for image processing and pattern recognition can make more effective the baggage screening process (He et al. 1997). A drawback for these kinds of devices, besides being very expensive, is that they are not very fast thus creating a serious bottleneck in the baggage flow. The development of high-speed devices with capabilities of comprehensive scanning is therefore crucial to ensure 100% baggage screening without disruption of the baggage flow. 32

33 Positive Passenger Baggage Match The basic idea behind PPBM procedure, as a supplement to baggage screening activity, is not to allow an already checked it in baggage to be transferred on a passenger commercial flight unaccompanied by its owner. Checked in passengers and their baggage s must be on the same flight. This baggage reconciliation policy is in effect for all flights in Europe and for international flights in North America (Jackson, 1999). The PPBM process activity does not take place until the moment the aircraft doors are closed (IATA, 1999). At that time after the cabin crew s crosschecks, between the numbers of passengers on the boarding list to the actually boarded ones, if a passenger has checked in baggages but not boarded on the same airplane then the appropriate baggages must be taken out of the plane. The task to locate unaccompanied baggage s for unloading is not an easy one as baggages are randomly arranged in the aircraft s containers. The actual procedure followed is to unload all the baggages for reconciliation by the passengers. Procedures that can take quite some time to be completed and thus imposing additional costs to airlines by delaying the flight. In an effort to overcome or alleviate this problem several IT based methods are applied or are under development. A proposed solution that could reduce the time required for PPBM procedure is to link Radio Frequency Identification (RFID) bag tracking systems to smart boarding cards (Drury, 1999). This way bags are processed for loading on to the aircraft only after there is confirmation that the owner has showed up for boarding on to the aircraft. RFID can also be used for fast retrieval of individual bags from the aircraft containers with the utilization of laser arrays (Yfantis, 1997). It can be accomplished by storing characteristics of bags into computers and assigning specific location in the aircraft s containers. At New York JFK International Terminal a comparable system is in use even though additional baggage storage area is required (Airport World, 1998). 33

34 Post September 11 Process Requirements Following the tragic events of September 11, 2001 among other measures increased baggage screening requirements have been set internationally. Governmental and air industry agencies require that 100% of the passenger baggage s must pass through screening devices compared to random checks prior to September 11. Consequences of these additional baggage-screening requirements are increased process times and a higher cost for additional faster and more accurate screening devices. The practice being departing passenger s baggage-screening process to take place after the check-in procedure in an area not accessible to the passengers, after the need for 100% baggage screening it could be more efficient to exercise baggage control prior to check-in in some airports. This practice may add problems to terminal s passenger capacity, as additional area is needed for queues in front of the screening devices, but could enhance the safety of baggage screening processes. If baggage-screening processes take place at the presence of the passenger this could act as a deterrent for any possible misplacement of potentially dangerous objects inside baggages. These additional baggage-screening requirements have also a negative effect on some previous practices with off airport baggage check-in activities due to the even higher costs involved. Nowadays, only a few cases of this kind of practices are implemented, in Switzerland and Hong Kong Check-in Process Among the airport terminal processing activities for departing passengers, including ticketing, baggage control, security screening, passport control and boarding, the check-in and baggage drop activity is considered the most important process. Passengers go through the check-in process to acknowledge their presence at the airport and their intention to board the flight they have made arrangements for. Their ids and other necessary travelling documents are verified. Then, they are allocated to a seat after presenting to the check-in staff a valid ticket. Seat allocation could take some time particularly with 34

35 passengers with special requirements such as passengers with reduced mobility, unaccompanied minors, as well as last minutes travellers on a waiting least. Next passenger baggages are dropped off on a conveyor, after they are weighed so as to comply with per traveller weight limitations imposed by the airline, to be transferred to the aircraft. They are labelled with an identification tag containing all the necessary information such as destination and flight number, which information are also saved on a database in each passenger s file. A baggage sticker is handed over to the passenger to be able reclaim them at the destination airport. Finally, when all the check-in activities are completed a boarding pass is issued to the passenger containing information about passenger s name, flight number, departure time and exit boarding gate. With the use of the boarding pass the passenger will be allowed to access restricted terminal areas after going through security checks in order to board the airplane. It is obvious that the check-in process takes a considerable amount of time to perform all these activities. The waiting time for available check-in counters accounts for more than 80 percents of all waiting time in the terminal area (Takakuwa and Oyama, 2003). It is these characteristics of the check-in process that makes it critical in the overall design and operation of the airport terminal activities. A most recent IT based development, utilized as a supplement to ordinary airline check-in counters and can greatly reduce values of passenger process time, is the operation check-in kiosk systems Check-in Kiosks From the airport terminal activities the one that has the greatest potential for improvements due to its complexity is the check-in process. For that purpose advances and innovations in IT such as the Common Use Self-Service (CUSS) kiosk systems were used to improve check-in process efficiency. One of the first attempts to improve the check-in and other terminal processes was the common-use terminal equipment (CUTE) initiated by SITA in Airlines using this system share counters that have computer terminals 35

36 connected to a common central computer system. CUTE systems gave airports and airlines more operating flexibility and considerable savings (Feldman, 1999). A recent extension to CUTE systems is the usage of CUSS kiosk that can be located and operate not only inside the airport terminal but also, with the adoption of wireless IT, at remote locations (IATA, 1995). The first CUSS kiosk systems were installed by individual airlines and could be used only by their own passengers. The world s first IATA-approved CUSS kiosk systems that were installed in Canada s Vancouver International and Tokyo Narita in late 2002 allow passengers to check-in regardless of which airline they are flying with (ACI, 2004). These kiosks are installed and operated by the airports themselves and not by individual airlines. The CUSS kiosk system enables passengers to check-in faster for a flight from a remote location but faces some problems with remote baggage drop activity. These problems have mainly to do with the issuing of the baggage sticker and the implementation of weight and size limitations. Very few cases of remote baggage drop systems under operation can be found. In Europe the first installed CUSS kiosks with integrated baggage transportation facility were at Vienna International in August 2002 with the first trial taking place a year later on October 2003 (ACI, 2004). The use of CUSS kiosk systems seems to be more suitable for passengers carrying handbags only. They can complete at a remote location all the required check-in activities get a boarding pass for their flight and pass through security processes at the airport terminal later. Especially after the events of September 11, 2001 the demand for enhanced security in identification processes complicated even more some check-in activities causing more passenger inconvenience. The usage of IT and new technologies shifted to focus more on enhanced security rather than on to reduce congestion, improve operational efficiencies and reduce costs. The need for balance between efficiency and security forced airport authorities and airlines to search for new solutions and methods. The air industry has to standardize these methods globally through air travel organizations like IATA so as to cut cost, increase efficiency and enhance security. 36

37 2.2.4 Security Checks There are two basic processes that a passenger has to go through before entering the secure area of the departing hall. Boarding pass control, which is a check of passenger s travel documents, and security checks with the use of metal detectors and X-ray systems. Boarding pass control is carried out to ensure that the passenger has a valid boarding pass for a specific flight and should be admitted into the secure terminal area. Furthermore, by checking that the name on the boarding pass matches that in an authorized Id document, to confirm that the passenger is the same person to whom the boarding pass was issued. The passenger before entering the secure area of departing hall has to go through a metal detection system and process any carry-on bags through an X-ray machine for screening. The purpose of this security process is to ensure that the passenger does not pass explosives, weapons or other potentially dangerous objects into the secure area of the terminal building. Some times additional hand search is conducted to randomly selected passengers to enhance even more security. The security checks are a significant factor on the boarding process and it may become a major cause o delays by creating bottlenecks. The processing time of security checks determines the arrival rate of passengers at the departure hall (Wirasinghe and Shehata, 1988). Increasing throughput by operating more advanced and accurate screening devices will reduce processing time and make the process more efficient. Keeping an optimum balance between the numbers of operated devices and the number of the passengers to be processed also affects processing time Customs and Immigration International passengers, in addition to the previous mentioned processes carried-on by domestic passengers, are required to go through immigration process in order to board a flight. They must posses valid travel documents, passports and visas for example, in order to ensure admissibility to their destination country. Further immigration and quarantine checks with state immigration control officials may be required in some instances. In European 37

38 Union, according to Schengen Convention for passengers that fly to states that are full members of the Schengen group there are no longer any border controls. The Schengen Agreement in 1985 and the subsequent Schengen Convention in 1995 abolished controls on internal borders between the signatory countries Boarding Boarding process is the last process that a passenger has to go through before leaving the terminal area to board on the aircraft. At the exit gate, after passenger displays the boarding pass, airline staff verifies the Id of the person and acknowledges that he or she will board the right flight. Also through the airline system confirms that the passenger is on board and transmits that information to baggage handling for the purpose of PPBM. 2.3 Simplifying Passenger Travel Program The SPT program is a join initiative of companies and organizations representing airlines, airports, travel agents, passengers, technology suppliers and broad government interests in order to expedite passenger processes. Some of the benefits for the parties involved and mainly for travellers are: reduced waiting and processing times, increased security through biometric identification and reduced costs through more efficient passenger processing. In 1997, IATA at its Annual General Meeting passed a resolution on Passenger Friendly Flows through Airports in order to improve the passenger experience globally (Durante, 2004). The initiative joined by ACI, AUC (Air Transport Users Council), ICAO, WCO (World Customs Organisation) and they formed the MSSG (Multi-Sectoral Steering Group). This was an initiative called the Simplifying Passenger Travel Program. The SPT s basic vision is of a one-stop identification check, instead of repetitive id checks, at the point of departure. It will be based on smart cards held by the passengers that contain machine-readable API (Advanced Passenger Information)/PRN (passenger Name Record) biometric data as well as passport and visa information. All these information s will be shared electronically among all service providers. This way, government required 38

39 data from the airlines concerning passengers would be collected in advance and not at the terminal check-in area. Passengers through an on-line system or via Internet would provide the data. These would allow processing to be done in advance of departure thus reducing processing time and enhancing security. In order to ensure the integrity of the gathered API/PNR data an independent authority must be formed through legislation to facilitate the entire process. There are privacy concerns with regard to stored and shared API data that must be addressed. Also in order these data can be collected automatically, globally agreed standards on the contents and format of the required data must be followed. This will also make the process faster and more accurate. Other technologies that could be used for passenger identification or access control in an airport terminal include high-tech devices that could recognize irises, fingerprints, hand geometry, facial characteristics and voices. Iris recognition is already used in several trials, and technology vendors are proposing various innovative approaches. In Europe, is required that the Member State governments adopt a harmonized approach in the implementation of the security issues. The cost of implementing all security measures must be covered by national governments and must guarantee the highest level of security and privacy possible. The USA has legislated, that biometrics data must be included in USA and visa waiver countries passports by October 2004 (Durante, 2004). The USA Transportation Security Administration (TSA) has conducted trials of various biometric technologies at more than 20 airports over the past two years. Japan conducted on 2003 a three-month electronic check-in trial by allowing Japan Airlines frequent flyers to register facial and iris images in order to experience faster e-check-in, security check and boarding through biometric verification (Durante, 2004). In order for all these innovations to be functional, in support to current practices in passenger processing, all interested parties must perform further trials and also familiarize passengers on the use of innovative technologies. Applying state of the art technologies on required prerecorded passenger data and passenger image recognition waiting and processing time will be reduced, security through biometric identification will be increased and 39

40 passenger processing will be more efficient at reduced costs. This way air industry will move forward to meet the demand for changing travel trends and growing traffic. 40

41 CHAPTER 3 Research Methodology and Data In this section the methodology that will be followed to utilize effectively existing airport terminal passenger facilities in order to manage congestion in a cost efficient way will be presented. Also description of the parameters and the data and data sources referenced will be provided. The methodology will be based on the design of a basic airport terminal model and the simulation of passenger processes and flows through the model s passenger facilities. Several different simulation scenarios will be implemented by altering model s options such as: Passenger facility configuration Input model data Parameters (attributes) of the model facility counters Running the model for different simulation scenarios output statistics reports including values on passenger process service and wait (queue) time will be generated. Based on the analysis of the output statistics reports and charts the type of options that significantly affect passenger processes and flows through the model s facilities will be specified. The final set of simulation scenarios based on the specified options will be implemented. The time values derived from output statistics reports of the final set of simulation scenarios, will be fetched into excel spreadsheets with the appropriate formulas and data on cost values for the calculation of selected types of costs. Economic analysis, based on waiting line theory, of the calculated costs will be conducted. 3.1 Simulation Model Methodology Operational efficiency is a crucial objective for airport and airline managers when they allocate and/or use the available facilities in the airport terminal area. Terminal facilities are fixed thus have capacity constrains on how many passengers can process during a specified period. Performance standards and predefined passenger level of service criteria, set by ICAO and other 41

42 organizations, must be employed. Many operations research techniques such as linear/integer programming, stochastic programming, and queuing theory can assist managers on utilizing these standards or modelling problems arising in airport terminal passenger processes, but they often fail due to poor scalability or excessive computational requirements (Hafizogullari et al. 2002). Instead discrete event simulation is often used to model systems where complex processes are combined with a limited infrastructure capacity such as in airport terminal passenger processes (Horonjeff, 1994). It is also used to analyse and improve airport terminal passenger processes and flows when the terminal capacity is near its limit (Babeliowsky, 1997; Joustra and Van Dijk, 2001). Not only airport terminals but also airports in general are ideal fields for applying simulation especially when some important inputs to analysis are unavailable (Horonjeff, 1994). Given that the flow of passengers through an airport terminal is a very complex dynamic process, as it is a function of time and space, discrete event simulation modelling methods will be employed. It will enable us to: Represent terminal facilities-counters by a set of unique conceptual elements that will be used to simulate their real-world behaviour. Model and study the interactive effects between most of the resources that handle passenger flows inside the airport terminal. Use any set of probability distributions for passenger arrival-departure patterns and terminal facility-counters passenger processing times. Implement suitable what-if scenarios for the number of terminal facilitiescounters in operation and passenger processes followed. Visual animate terminal facilities-counters and passenger flows inside the model terminal. Even though discrete event simulation seems to be a good research instrument for analysis of airport terminal passenger processes, the building of the model must be concise in order to be effective (Valentin, 2003). Otherwise wrong results could be drawn that could lead to ineffective solutions during real world implementation (Keller et al. 1991). The simulation model will focus on departing passenger processes inside the terminal building, as they are more complex and time consuming compared to 42

43 arrival processes, and have a greater effect on delays and associated costs. The stages of simulation processes that will be followed are shown in Figure 3.1 (Render, 2000). Define Problem Introduce Important Variables Construct Simulation Model Specify Values of Variables to Be Tested Conduct the Simulation Examine the Results Select Best Course of Action Figure 3.1 Simulation process (Source Render, 2000) Simulation Model Software Some of the existing simulation tools for the development of airport terminal passenger models (Render, 2000; THENA, 2002) are: Passenger SIMulator (PAX-SIM) Pand Baggage Flow Model (PBFM) SIMMOD (SIMulation MODel) Simple Landside Aggregate Model (SLAM) Total Airport Airspace Modeller (TAAM) WITNESS Witness was selected as the software tool for the design of the basic airport terminal model and the simulation of passenger processes and flows through the model s passenger facilities as it is capable of simulating a variety of discrete and continuous events (Bouamra and Morrison, 2000). It offers great design flexibility as the model can be created dynamically on a display and 43

44 the designer can interact with the model elements while the program is running. The designer elements that graphically portraits the layout of the passenger terminal facilities such as check-in servers are drag-and-drop predefined objects onto the simulation screen. The elements can be in any number of states while the model is running such as idle (waiting), busy (processing) and blocked (stopped). The passenger flow processes such as queues are graphically displayed on the screen during the running of the simulation. The simulation can run in different modes, from step-by-step to a batched time in the future. Statistics reports and charts for the entire simulated process or any of its components are generated automatically in different formats. These reports will assist us to choose between alternative modeling scenarios or will be fetched into other software packages for additional analysis. Tables and color-coded histograms, time series and charts will be created for each terminal process for any specified period of time. The Witness simulation software will run on a computer with the following (basic) configuration: Intel P III Processor (386) Windows 98SE (95) 64MB RAM (8MB) 100MB free disk space (10MB) SVGA (800X colors) screen resolution (800X600) Model Assumptions The design of a system model requires simplifying assumptions to limit the complexity of the computer programme that needs to be included. This computer programme provides the necessary degree of logical controls over the processes of the simulation. It also needs to consider the speed and storage capacity of the computer system being used to run the simulation software. However, the model still must provide sufficient descriptive details of the terminal facilities and passenger processes for accurate simulation. The general assumptions in the design of our model are the following: 44

45 Passenger facilities and processes are in general similar and independent of airport terminal type. Basic type of facilities and passenger flows inside the terminal will be configured. Passengers proceed directly from facility to facility and wait for processing in front of any group of facilities in single queues on a first-come firstserved base. Each passenger service facility operates independently but they all form a chain of processes for the passengers, geared up by the previous one. The number of passengers in the model will depend on the type of aircraft being served and its loading factor. Common passenger profile for different class travellers (Business, economy, charter). International passengers, who are citizens of the Schengen group countries and have, as destination Schengen member states will be treated as domestic. All of the passengers are generated and are available for processing at the start of the simulation Model Passenger Flow Description The functional flow diagram of the departing passenger processes as can be seen in Figure 2.3 will be used as our conceptual model for the design and construction of the real model for our simulation study. The outline of the simulation model that was designed with all the appropriate terminal facility counters and passenger processes-flows can be seen in Figure 3.2. The Departure hall inside the designed model is where both domestic and international passengers will appear while they will be generated by the simulation software. From the passengers who will be generated a very small fraction (6%) of them will be led to the TICKETS counters to purchase a ticket. Passengers proceed then to the Bag. Control security systems and only passengers without baggagges (5%) proceed directly to the CHECKIN counters. After, Bag. Control checks they also proceed to the CHECKIN counters. The next step for all passengers domestic and international alike is 45

46 to go through X ray machines and then lead to W. Hall Dom and W. Hall Int. respectively. In addition international and non-schengen country travellers will pass the Passport control prior entering the W. Hall Int. area. This way the departing passenger processes inside the terminal building will be completed. Figure 3.2: Model facility counters and passenger flows Input Data Model Parameters In order to evaluate the different stages of the terminal passenger processes and to obtain valuable statistical outputs the following input data are required for the proposed model to run: Number of passengers according to aircraft type Loading factor of each aircraft type Type of flight (passengers) Domestic/International Percentage of passengers without a ticket Percentage of passengers without baggage 46

47 Service rate-distribution of processing at each facility Number of facilities at each type of passenger process Queue capacity in front of each facility (number of passengers waiting to be served) Queue delay time The required data mainly refer to the type of passengers being served, the seating capacity and loading factor of aircraft being served and the service rate of passengers at each terminal facility. The information were gathered from public sources, as historical data are available through Eurocontrol, FAA, IATA, airport web pages, on site data collection and the recommendations of ICAO and ACI. The aircraft type and its exact seat number sample that was selected are shown in Table 3.1 (Eurocontrol, 2004A). The sample was selected on the basis that these types of aircraft have the highest contribution to total ground movements in most European airports (Eurocontrol, 2004A). Seat capacity range covers minimum and maximum values, thus the results of our simulation will be valuable. In the sample was also included the new Airbus A for experimental purposes. Table 3.1: Aircraft Type Passenger Capacity (Source Eurocontrol, 2004A) Aircraft Type Seat Number B B B B B ER 240 A A A A A A ATR

48 The passenger load factor that is recommended by Eurocontrol for 2004 is 59.4% % (Eurocontrol, 2005). In our model the highest value of 80.5% will be used, as the purpose of our simulation is to examine the passenger processing capabilities of the terminal facilities during periods of demand. The passenger load factor represents the percentage of seating capacity that is sold and utilised on most of the flights with each of the above aircraft types. Data on service rate and distribution of processing time at each terminal facility are summarized in Table 3.2. The values of the table are in time units and the unit of time is a minute which means that for the ticketing process the actual three values for Min Mode and Max will be 1min, 1min and 12sec (1.20*60) and 1min and 18sec (1.30*60) respectively. The representation of the values has to be in this format as this is a Witness software constraint and has to be applied through out the model. The statistical data of the Table 3.2 that were obtained from on site collection and other sources such as ICAO (1999), Takakuwa and Oyama (2003), (Hafizogullari et al. 2002) and (Casado et al. 2004) since they portray variations it was decided to model them as triangular distributions (Kelton et al. 2001). The triangular distribution is typically used when statistical information on the area under study is hard to obtain, but the likely range of values and the most common value are known. Table 3.2: Data on Airport Terminal Facilities Terminal facility Processing time (min unit of time) (Triangular distribution) Min Mode Max Ticket counter Baggage control Check-in counter X ray control Passport control Percentage of passengers without a ticket and percentage of passengers without baggage will be set to 6% and 5% respectively as there are not available data on these values. Sensitivity analysis for these parameters will be contacted, with 95 % confidence interval, in order to examine if the above 48

49 values are critical for the whole terminal process by comparing the results of the various simulations. Queue capacity in front of ticket, baggage control, check-in, x ray and passport facilities will be set to 10,25,30,30, and 25 passengers respectively as these figures don t have an effect on delays in processing time of passengers but rather on the level of standards experienced by the passengers inside the terminal area. Queue delay time for q1 to q5 will be set to a range of values from 0.5min to 3min and will be tested if they are critical for the entire terminal passenger processes. Queue delay time denotes the minimum and maximum time spend by a passenger inside the queue (Buffer). Detailed Figures of the above model computer programme implemented data on facility-counters and passenger processes can be seen in Appendix B Model Implementation The model is implemented with the combination of Witness building bocks called modelling elements, which are created and then linked together with rules in order to simulate terminal facilities and the passenger flows through them. It is build and tested incrementally from one passenger process to the next so as to minimize arising problems and to assure the validity of the model. Also contains build-in information regarding the input and output rules that are used on key elements and an interactive microcomputer program that simulates all the processes of the terminal and gives the necessary degree of logical control linking the elements. Other information such as percentage of passengers without ticket and without baggage, the cycle times, distributions and actions of counters-elements and the capacities of the buffers-queues are also incorporated into the model. During the loading of the model simulation programme for implementation all the counters of the different facilities are activated except the check-in counters that must be activated, through an interact box, according to the scenario under examination. Also the second x ray and baggage control counters are programme activated if more than one hundred passengers will 49

50 proceed through them. The layout of the model with all the facility counterselements that was implemented is shown in Figure 3.2. Several different simulations will be implemented by altering input and model data to confirm that the simulation model is working as expected and that the results of the simulations are valid compared to existing standards for passenger terminal processes. Then in order to evaluate and compare different terminal facility configurations multiple independent scenarios will be implemented with the help of the simulation model. The results of these scenarios will be fetched into excel spreadsheets for cost analysis which will be the tool that will assist us on deciding optimal terminal facility configuration. 3.2 Economic Data on Cost Values System costs that will be examined in this dissertation include passenger value of time costs (VOT), airline terminal facility service costs and airline delay costs. Since detailed financial information of airlines are not all readily available due to confidentiality and competition, some cost values will be calculated approximately from publicly available financial data. Besides the aim of the proposed simulation model and of the succeeding cost calculations is to demonstrate rather than to precisely portray cost values of any specific airline in the industry, for analysis purposes. Nevertheless, this simplification in the calculation of costs does not weaken the capabilities of the proposed airport terminal model. If prospective users have accessibility to confidential airline information or retrieve the appropriate values on the detailed cost data, when they will be publicly available for analysis, they can utilize the output results of the model. Passenger value of time unit costs The average value of time per passenger is related to the average wage costs, to the purpose of the travel and the distribution of the traveling passengers profile, business or economy. It was estimated by ITA (2000) and is recommended by Eurocontrol (2005) in the range of per hour per passenger. In our model for simplicity calculation purposes VOT cost is assumed to be 49 per passenger per hour during waiting time at an airport terminal. 50

51 Airline terminal facility service unit costs Economic data on facilities service costs at the airline level are not readily available in a comprehensive way, as there are significant policy differences in cost evaluation among the airlines. Furthermore facility service costs can vary for the same airline at different airports as it depends on facility charging policies by each airport operator. Since there are significant differences in facility costs it was decided to correlate it with estimates of labor costs, for airline employees serving terminal passenger facilities, and thus to ensure the reliability of our model for analysis purposes. Estimates of labor costs can include direct salary or wage costs plus the cost of benefits but also the company s overhead costs that are allocated to labor as well as the direct cost and the cost of benefits. It was decided to assume a wage cost of 30 per hour per airline worker employed at the airline passenger terminal facilities for our simulation model calculations. Airline delay unit costs A range of different values on estimates for delay costs can be retrieved from existing studies. Differences on estimates of delay costs result mainly from the method that is applied by each study for the computation of the delay time and the type of costs that are included for the evaluation. It was estimated by the University of Westminster and is recommended by Eurocontrol that the average cost of ground delay to be 72 per minute (Eurocontrol, 2004A; Eurocontrol, 2004B). This value indicates the average cost per minute that the airline incurs of delaying a passenger and takes into account aircraft delayed, crew salaries and expenses, handling agent penalties and airport charges (Eurocontrol, 2004B). In our model for calculation purposes airline delay cost will be set to 72 per minute for delays of 15 minutes and above (Eurocontrol, 2005). 3.3 Cost Computation Formulas Waiting line theory has been the primary source for developing models to conduct economic analysis of airport terminal processes (Lemer, 1992). Waiting line problems in general focused about the question of finding the ideal level of services that an organization should provide. Figure 3.3 shows 51

52 the general shapes of costs curves in an economic analysis of waiting (queue) lines. From the shape of the curves is obvious that waiting cost decreases while service cost increases as the numbers of service channels (counters) are increased. The optimal service level (optimal number of counters) is where total cost is approximately minimized and can be found by evaluating the total cost for several design alternatives of the model (Render, 2000). Before an economic analysis of a waiting line models can be conducted, a total expected cost model, which includes the cost of waiting and the cost of service, must be developed. Figure 3.3: Waiting line cost curves For practical reasons, all types of cost computations will be based on the application of the previous decided unit cost values as shown in Table 3.3, to appropriate time measures derived from our simulation model. All values of Table 3.3 are given in euros at 2004 price levels and for any use in future years they must be adjusted according to appropriate price and income level changes. 52

53 Table 3.3: Unit cost values (Source Eurocontrol, 2005) Type of cost Value per minute ( ) VOT 49/60 Service 30/60 Delay 72 The main three types of cost which incur during passenger flow process through the airport terminal facilities under certain conditions will be calculated for evaluation purposes: Passenger value of time cost Airline terminal facility service cost Airline delay cost Each of these costs is and will be consider as a linear function of its time duration (Wu and Caves, 2000). Finally total cost will be calculated as the sum of the above individual costs. Passenger value of time costs The formula that will be applied to calculate passenger VOT costs based on waiting line -queuing theory- is: Total Passenger VOT Cost = VOT Unit Cost x Total Passenger Waiting Time Total passenger waiting time is the sum of time that all the passengers spent waiting to be processed inside the terminal building from the time they appear in front of the first processing facility till they arrive at the departure hall. Is the time they waist waiting in queues and moving from one facility to the next, and is estimated by our simulation model for each scenario under analysis. Airline terminal facility service cost From all the terminal processes, check-in process is the most critical time consuming terminal process and is the only process that can be controlled and influenced by the airline (BAA, 2001). Airline can influence the process time of this facility by operating the appropriate number of counters with the right staff level. Airline s terminal service cost is determined mainly by the 53

54 labour cost of the check-in counter personnel employed. The formula that will be applied to calculate airline terminal facility service cost is: Total Service Cost = Service Unit Cost x Number of Check-in x Service Time Number of Check-in is the quantity of assigned check-in counters in our simulation model for each alternative scenario under examination. Airline delay cost Airline delays will occur if the total time spend to process all passengers of a certain flight exceeds the recommended by ICAO (1999) 60 minutes practice. As total time (Process Time) spend is defined the time elapsed for the entire passenger flow process to be completed for a certain number of passengers. The formula that will be applied to calculate airline delay cost is: Total Delay Cost = Delay Unit Cost x ( Passenger Flow Process Time) Our simulation model for each alternative scenario under examination, will calculate the values of the passenger flow Process Time. To the ICAO 60 minutes recommended value we will add 15 minutes, as it is realistic to apply zero cost to delays lower than 15 minutes (Eurocontrol, 2005). Airlines tend to integrate 15 minutes buffers into their scheduled departure times for tactical reasons. Finally total cost, as the sum of passenger VOT cost and airline service cost will be calculated. All the calculations of the different types of costs will be done with the help of excel spreadsheets where all the different time values for each experimental scenario will be fetched automatically by our simulation model. The analysis of the cost outputs and their corresponding plots will assist us to decide on the most effective terminal facility configuration for efficient passenger processing. 54

55 CHAPTER 4 Results and Evaluation In this section the overall results of the implemented terminal simulation model and relevant calculated costs will be presented and evaluated. Simulation outputs generated counts of facility status, waiting line and queue length statistics for all the alternative scenarios of terminal facility configuration performed. Time dependent performance characteristics of passenger flows derived from the implemented model scenarios will be fetched as input data for cost calculations. Economic analysis of the relevant costs calculated for each simulation scenario will be conducted in order to determine the most effective configuration of the terminal facilities for cost efficient passenger processing. Cost efficient passenger processing will assist us to alleviate terminal congestion and consequent costly delays. Figure 4.1: Terminal passenger flow 55

56 Furthermore, as can be seen in Figure 4.1, the visual animation of passenger flows through the terminal facilities also provided valuable information for our study purposes. 4.1 Simulation Results We performed several alternative simulation scenarios on the terminal model we designed and obtained a huge amount of outputs. The simulation outputs contained detailed activity statistics reports and charts for all the alternative scenarios we run on the following: Number of processed passengers Average process time of the entire simulated terminal process Type of flight Domestic (D) or International (I) Queue names and numbers Number of passengers in every queue Queue sizes and time Name and number of counters of the terminal facilities Counter utilization Number of operations per facility counter (Number of passengers served) Conveyor (Passenger walkway) size and utilization. Percentage of time spend in front of each facility s queue Percentage of total terminal process time spend per passenger state Percentage of total service time spend per terminal facility All these valuable simulation statistical results were written to different types of output files continuously and portrayed to related charts. For demonstration purposes simulation statistics reports and charts, for one of the scenarios we implemented and are shown in Table 4.1, appear in Appendix C. Table 4.1: Demonstration Simulation Scenario Flight Type Number of Passengers Number of Check-in I

57 Studying the output statistics and charts was decided to run alternative scenarios of the simulation model by concentrating on the options that significantly affect the average process time. These options are the following: Flight Type Number of passengers Number of check-in counters For some of the other options and alternative model parameters such as number of ticket counters, percentage of passengers without ticket or baggage, queue capacity and queue delay time was evident from the simulation results that their values don t influence the total terminal process time to a great extend. Terminal process time is the time needed by each alternative simulation scenario to be completed and measured by our model. Before the detailed examination of the main simulated scenarios results a brief outlook of the findings, only for the option of passenger without ticket and baggage, will be presented in support to the above statement Without Ticket-Baggage Passenger Scenarios Running the simulation model for different percentages of passengers without ticket and baggagges it became evident that these values are no critical for the terminal process. They affect very slightly the terminal process time as can be seen in Table 4.2, thus not influencing our final results. Table 4.2: Without Ticket-Baggage Scenario Passengers Without Ticket (%) Passengers 140 Passengers Without Baggage (%) Average Process Time (minutes) Check-in 3 Check-in The corresponding values for the base case scenario without ticket-baggage were set to 6% and 5% respectively as was defined in section Executing sensitivity analysis on each of this values using normal distribution 57

58 (sd=1) it was found that for 95 % probability the observations would approximately fall within the range of values shown in Table 4.2. We decided to run only eight simulations using the corresponding values, instead of eighteen (matrix 3x3x2), which is the correct method. This was decided and is self-evident that the last pair of values (8%-3%) would be the most affecting on the average process time, giving the highest value. The selected value of 140 for the number of passengers was based on Appendix D Table 1 as this value corresponds to the average figure of aircraft seating capacity for the given aircraft types in Europe (Eurocontrol, 2004B). We used three and four check-in counters to confirm that the influence of the number of check-in counters is much greater than the number of passengers without ticket or baggagges on the process time value measured by our model Main Simulation Model Results We implemented several different simulation scenarios for each of the options mentioned in section 4.1. The total number of scenario cases we had to run is the multiple of the number of different cases for each of the three selected options. For the first option we had to run the model under two different passenger profile cases (Flight Type), Domestic-International. In order to implement this option we had to run the model under two different terminal processes, one for Domestic and one for International. For the second option we had to run the model under different number of passengers, twelve cases as can be seen in Appendix D Table 1, by generating each time the appropriate number of passengers. For the third option we had to run the model under different number of checkin counters, one to ten cases, by activating each time the appropriate number of counters. The total number of scenarios we run was (2x12x10) two hundred and forty and the results can be seen in Appendix D Table 2 and Table 4. 58

59 Model Results Analysis Tables 2 and 4 in Appendix D provide performance characteristics of the passenger flow through the terminal facilities, which are: Process Time Service Time Blocked Time Walk Time Queue Time Process Time is the total time estimated by the simulation model for the entire terminal process to be completed under each case of flight type, number of passengers and number of check-in counters. Service Time is the time estimated by the model for all the passengers of each case to be serviced by the appropriate number of the check-in counters airline personnel. The sum of Blocked, Walk and Queue Time is the total passenger waiting time estimated by the model for the entire terminal process to be completed under each case of flight type, number of passengers and number of check-in counters. From the animation of the passengers flow through the terminal we visualize the points where passenger blockage-waiting occurs or queues are formed as can be seen in Appendix E Figure 1. For example for the case of 140 passengers and one check-in counter active we observe that passengers are blocked at the baggage control facility and queues are formed in front of the facility. Moreover from the output chart of the Witness simulation for the baggage control, as can be seen in Appendix E Figure 2, activity statistics are shown for the facility. For the total time of minutes it takes for the passenger flow process to be completed for the case under examination, the baggage control facility almost 25% of the time is in idle condition, 8% is in busy condition and 67% is blocked. Result experiments of the above case are also shown in histograms of Appendix E Figures 3, 4, 5 and 6. In Figure E.3 we see that the utilization factor for the check-in facility is 97.8 %, which is percentage of time that the 59

60 check-in counter personnel were working. Figure E.4 shows that the total time spend by the passengers in queues, 0.58 % was spend in front of ticket facility, % in front of baggage control facility, % in front of the check-in facility, 1.43% in front of the x ray facility, 0 % in front of the passport facility (Domestic flight type) and % inside the Domestic Hall before the exit gate for boarding the aircraft. Figure E.5 shows that of the total time spend by the passengers % was spend for service receiving, % of the time passengers were blocked-waiting for the next available facility in order to be served, % waiting in a queue and 0.05 % of the time walking. Figure E.6 shows that of the total time spend by the passengers for service receiving % was spend at the check-in counter, % at the baggage control facility, % at the x ray control and 0.08 % at the ticket facility. Examining the results for the same number of passengers (140) but with different number of active check-in counters (3 instead of 1) we see from Appendix E Figure 7 that % of the time was spend for service and only % of the time passengers were blocked, compared to % with the case with one active check-in. The activation of two more check-in counters considerably reduces not only the process time from to minutes but also the time passengers spend waiting during the terminal process. Similar performance indicators can be analysed, for all the different scenarios we implemented, based on the output files produced by the Witness software simulation model. In Appendix C can be seen all the relevant output files, charts and histograms for the International Flight Type of 140 Passengers and 3 active Check-in counters. From the overview of graphs and histograms it is obvious that bottlenecks of the passenger flow process can be resolved by the utilization of additional facility servers, with the check-in facility affecting the most the value of time required for the entire process to be completed. Based on the results of the different scenarios we run, decisions can be made on the number of facilities that should be employed under each case. Even further on the next step, where the economic results will be analysed, the exact number of check-in counters required for efficient passenger flow inside the terminal building for each scenario will be determined. 60

61 4.2 Economic Results Analysis In order to estimate the optimal number of check-in counters under each scenario the different values for passenger service and waiting time that were calculated by the simulation model were fetched into excel spreadsheets for cost calculations and analysis. Tables 3 and 5 in Appendix D summarise the results of the economic cost calculations for each scenario. To make this evaluation, identify the optimal number of check-in passenger service counters, we have to decide on the balancing point between the costs incurred by the passengers and the service provider, airline in our case, for the different terminal facility layouts under each scenario. One cost is the VOT cost incurred by the passengers and the other the Service cost the airline has to pay for the personnel employed at the service counters. Moreover, for the departing passenger process to be efficient, must comply by the standards set by ICAO and airlines such as 60 minutes process time plus the buffer time incorporated by the airlines into their schedule (ICAO, 1999). Exceeding this value of process time additional costs are incurred by the airline as delay costs, which in general are much higher than the Service costs and must be taken into account for determining the optimal facility layout. Also as can be seen from Appendix D Tables 3 and 5 delay costs are increased with increasing aircraft size (higher passenger seat number) due to its higher operating costs. Based on economic analysis of waiting line theory charts are drawn, as can be seen in Appendix D Figures 1 to 12, for the Domestic type flight the appropriate set of Number of passengers and the associated active Number of check-in counters. Study of these charts will assist us to drawn some important conclusions on the most cost efficient service layout. These charts portray the trade-off between the different types of costs - process time and the number of active check-in counters. The trade-off being that, as for our study, to decrease or increase processing time or the different types of costs to desired levels of efficiency by employing the appropriate number of checkin counters. 61

62 The chart of Figure 1 in Appendix D illustrates the shapes of the curves for the VOT Cost, Service Cost, Total Cost, Total plus Delay Cost and Process Time for the case of 52 Number of Passengers for Domestic Type of fight for 1 to 10 Number of Check-in counters. The VOT Cost is almost linear and tends to fall as the number of active counters is increased but remains stable when more than 6 counters are employed. The Service Cost is linear and increases as the number of counters increases. The curve at the top is the sum of the above two types of costs, it starts out high and ends up high after passing through a minimum at the point where the two costs intersect. This is the point where the total cost is minimized and occurs when two check-in counters are employed. Therefore the optimal number of check-in counters for cost efficient passenger processing under this scenario is 2. Also the value of Process Time ( minutes) agrees with performance standards set by ICAO and airlines. Under this scenario we don t have any Delay Costs and as a result the curves of Total Cost and Total plus Delay Cost coincide as can be seen in the chart of Figure 1 in Appendix D. In the chart of Figure 7 in Appendix D that illustrates the curves for the 140 Passenger Number scenario delay cost must be taken into account in determining the optimal number of check-in counters. From the cost curves is obvious that the optimal number of check-in counters is 3 where not only Total Cost is minimized but also happens that Total plus Delay Cost is minimized. On the contrary in Figure D.12 where we have the cost curves for 528 passengers Total Cost and Total plus Delay Cost don t minimize for the same number of check-in counters. Total Cost minimization occurs when 3 check-in counters are employed but Total plus Delay Cost minimization takes place when 8 counters are in use. The most cost efficient passenger processing takes place with the utilization of 8 check-in counters, which resulted to a value for Process Time of minutes. Similar economic result analysis can be applied for all the different scenarios we implemented for either Domestic or International Flight type, in order to decide on the optimal number of active check-in counters for cost efficient passenger processing. Tables 4.3 and 4.4 summarize the economic result analysis indicating cost efficient values on the last column, shaded in grey. 62

63 The Tables also illustrate where Total Cost and Total plus Delay Cost coincide for the optimal number of check-in counters under each scenario. Table 4.3: Cost efficiency values-optimal Check-in number (Domestic) Α/Α Flight Type Number of Passengers Number of Check-in Process Time Service Cost VOT Cost Total Cost Delay Cost Total plus Delay Cost 1 D D D D D D D D D D D D D D D Table 4.4: Cost efficiency values-optimal Check-in number (International) Α/Α Flight Type Number of Passengers Number of Check-in Process Time Service Cost VOT Cost Total Cost Delay Cost Total plus Delay Cost 1 I I I I I I I I I I I I I I I

64 4.3 Model Implementation and Study Limitations The version of the Witness software package used for simulating the flow process of passengers inside the airport terminal building cannot run under Windows 2000 or Windows XP environment. A newer software version of Witness is certain that it will be able to run under these operating systems or additional ones like UNIX. The model that was designed could be integrated -as a building block- into a more complex model because of its flexibility. This way additional performance characteristic of the passenger flow process could be evaluated which could lead to more concrete cost efficient decisions. The simulation model results are sensitive to input data and parameter selection. If we had selected different input data or fewer output parameters for evaluation we may had obtained different results. For example, we could also evaluate the time needed by the passengers to board the aircraft or assume the utilization of check-in kiosks with very low service times. Nevertheless the simulation model designed and the results obtained portray analytically the departing passenger flow process and the costs associated with it. Incorporating the appropriate data and parameters (the actual one for each case under study) by interested researchers the results that will be obtained can assist to make the right decisions concerning performance standards and cost efficiency. 64

65 CHAPTER 5 Conclusions The continuing growth of air transportation frequently challenges the service quality and efficiency of the airport terminal processes. Departing passenger procedures such as, ticketing, security screening, check-in, passport control and boarding, often generate queues inside the terminal area. Waiting lines of airline customers, in front a terminal counter, are not unusual especially during peak hours of demand. Congestion and delays are created and as a result not only the passengers are dissatisfied as they waste valuable time in queues, but also airports and airlines incur economic cost and lose revenue. In order to alleviate these problems and operate efficiently different approaches exist. The long-term approach -construct new airports or expand existing infrastructure- seems to be the obvious way to increase capacity of airports in order to accommodate future increase in air passenger traffic, but requires huge capital investments. A more cost effective approach that can be realized in the short-run is to improve the way existing airport terminal facilities are operated at congested airports. In this dissertation we presented a study based on this approach, which assisted us to notice the following: The critical parameters that mostly affect the terminal s process time are the number of passengers being served and the number of check-in counters. By increasing the number of the check-in facility counters the process time of the passengers can be reduced. Passenger waiting times as well as average queue length values increase when aircrafts with large number of seat capacity are to be serviced. It is clear that, for a given number of passenger service facilities, service quality is likely to decline as the volume of passenger increases. Passengers spend most of their time waiting to be served and more than 70 percent of that time is spend for the check-in counter facilities. Delay Cost values are much higher than Total Cost values calculated for the entire process of serving the passengers inside the terminal. 65

66 The values of the different costs calculated depend on the seating capacity of aircrafts being served, the number of facility counters and increase accordingly. Decisions on determining the most cost efficient layout of facilities employed must be based on the Total Cost calculated for aircrafts with seating capacity of up to 193 passengers for the sample we used but for higher seating capacity must be based on the sum of Total plus Delay Cost values. Clearly, an optimal balance exists between terminal facility service rates, queues and delays, such that the sum of the costs involved is minimised and yet maintain a desirable level of service quality for passengers. Improvements can be made to the entire terminal passenger processes for cost efficient passenger handling, which could yield to potential savings for passengers, airlines and airports. 5.1 Recommendations for Future Research Although the alternative simulation scenarios implemented in this research clearly show the significance of the selected terminal facility configuration changes and passenger process attribute value changes on cost efficient passenger handling, the evaluation of the additional following issues could even further assist to accommodate future increased demand for airport passenger traffic. The employment of additional supporting staff at some terminal facility counters, which could have an effect on passenger process time and service cost. The use of a different type of facility configuration or the dynamic opening and closing of facility counters according to passenger profile such as business, economy or charter class traveller. The incorporation of innovative high technology such as Advanced Passenger Information and Machine Readable Travel Documents into passenger process facilities, although data on service rates of this type of facilities are not yet widely available. The effects on passenger process time and airline service cost of the facility counters equipment reliability, performance and cost. 66

67 APPENDIX A: Standard IATA Delay Codes (Annex 3) SOURCE: Provisional list composed by IATA Passenger and Baggage 11 (PD) LATE CHECK-IN, acceptance after deadline 12 (PL) LATE CHECK-IN, congestions in check-in area 13 (PE) CHECK-IN ERROR, passenger and baggage 14 (PO) OVERSALES, booking errors 15 (PH) BOARDING, discrepancies and paging, missing checked-in passenger 16 (PS) COMMERCIAL PUBLICITY/PASSENGER CONVENIENCE, VIP, press, ground meals and missing personal items 17 (PC) CATERING ORDER, late or incorrect order given to supplier 18 (PB) BAGGAGE PROCESSING, sorting etc. AIRPORT AND GOVERNMENTAL AUTHORITIES 85 (AS) MANDATORY SECURITY 86 (AG) IMMIGRATION, CUSTOMS, HEALTH 87 (AF) AIRPORT FACILITIES, parking stands, ramp congestion, lighting, buildings, gate limitations, etc. 88 (AD) RESTRICTIONS AT AIRPORT OF DESTINATION, airport and/or runway closed due to obstruction, industrial action, staff shortage, political unrest, noise abatement, night curfew, special flights 89 (AM) RESTRICTIONS AT AIRPORT OF DEPARTURE WITH OR WITHOUT ATFM RESTRICTIONS, including Air Traffic Services, start-up and pushback, airport and/or runway closed due to obstruction or weather, industrial action, staff shortage, political unrest, noise abatement, night curfew, special flights 67

68 APPENDIX B: Witness Model Activity Data Layout Figure 1: Model Title Figure 2: Queue Process 68

69 Figure 3: Ticket Counter Figure 4: Baggage Control Counter 69

70 Figure 5: Check-in Counter Figure 6: X ray Counter 70

71 Figure 7: Passport Counter 71

72 APPENDIX C: Witness Demo Outputs Statistics Reports and Charts by Total Simulation Time Table 1: Entity Statistics Report Name Number Entered Number Shipped Av. Time DepPassengers Name Table 2: Queue Statistics Report Total In Total Out Now In Max Min Average Size Average Time HDomestic HInternational Avg. Delay Count Avg. Delay Time q q q q q q

73 Figure 1: Queue Statistics Chart 73

74 Name %Idle %Cycle - Busy Table 3: Activity Statistics Report %Cycle - Filling %Cycle - Emptying %Stopped - Blocked %Waiting - Setup %Stopped - Down %Waiting - Cycle Number of Ops. Baggagecontrol Baggagecontrol chekin chekin chekin chekin chekin chekin chekin chekin chekin chekinc Passport (1) Passport (2) TICKET TICKET TICKET TICKET TICKET TICKET Xray Xray

75 Figure 2: Activity Statistics Chart 75

76 Figure 3: Activity Statistics Chart Status 76

77 Table 4: Conveyor Statistics Report Name %Empty %Move %Block %Queue %Down Now On Total On Av. Size Av. Time Off Shift d d d d d d

78 Figure 4: Conveyor Statistics Chart 78

79 Figure 5: Conveyor Statistics Chart Status 79

80 Figure 6: Check-in Utilization (%) Figure 7: Total Time in Queues per Facility (%) 80

81 Figure 8: Total Time Spend per Passenger State (%) Figure 9: Service Time per Facility (%) 81

82 APPENDIX D: Main Simulation and Cost Calculations Table 1: Passenger number per aircraft type A/C Seats Pass. Load Factor Boarding Pass. ATR % 52 B % 81 A % 101 B % 115 A % 125 A % 134 B % 140 B % 175 B ER % 193 A % 233 A % 302 A % 528 A/A Flight Type Table 2: Domestic Flight Type Scenario Cases Number Number of of Checkin Passengers Process Time Service Time Blocked Time Walk Time Queue Time Delay Time 1 D D D D D D D D D D D D D D D D D D D D D

83 22 D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D

84 67 D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D

85 112 D D D D D D D D D Table 3: Domestic Flight Type Cost Calculations Service Cost VOT Cost Total Cost Delay Cost Total plus Delay Cost

86

87

88 Table 4: International Flight Type Scenario Cases Α/Α Flight Type Number of Number of Passengers Check-in Process Time Service Time Blocked Time Walk Time Queue Time Delay Time 1 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I

89 36 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I

90 79 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I

91 Table 5: International Flight Type Cost Calculations Service Cost VOT Cost Total Cost Delay Cost Total plus Delay Cost

92

93

94 350,00 300,00 250,00 200,00 DP ,00 100,00 50,00 0, Total Cost 106,49 94,41 110,39 134,64 161,01 185,96 211,79 238,37 264,94 291,52 Process Time 55, , , , , , , , , ,3224 Service Cost 26,32 52,44 78,59 105,82 134,06 160,16 186,05 212,63 239,20 265,78 VOT Cost 80,18 41,97 31,80 28,83 26,95 25,80 25,74 25,74 25,74 25,74 Total plus Delay Cost 106,49 94,41 110,39 134,64 161,01 185,96 211,79 238,37 264,94 291,52 Figure 1: Number of Passengers , ,00 800,00 DP ,00 400,00 200,00 0, Total Cost 233,97 168,81 180,71 210,17 247,99 286,22 325,75 366,40 407,05 447,69 Process Time 85, , , , , , , , , ,8987 Service Cost 41,13 81,61 122,69 164,45 205,23 245,20 284,54 325,19 365,84 406,49 VOT Cost 192,84 87,21 58,02 45,73 42,76 41,02 41,21 41,21 41,21 41,21 Total plus Delay Cost 984,98 168,81 180,71 210,17 247,99 286,22 325,75 366,40 407,05 447,69 Figure 2: Number of Passengers 81 94

95 3000, , , ,00 DP ,00 500,00 0, Total Cost 339,38 237,97 240,19 264,51 310,94 360,03 411,74 462,89 514,05 565,20 Process Time 105, , , , , , , , , ,2276 Service Cost 51,01 102,08 153,76 202,93 255,47 306,58 358,08 409,24 460,39 511,55 VOT Cost 288,37 135,89 86,43 61,58 55,47 53,45 53,66 53,66 53,66 53,66 Total plus Delay Cost 2514,06 237,97 240,19 264,51 310,94 360,03 411,74 462,89 514,05 565,20 Figure 3: Number of Passengers , , , , ,00 DP , ,00 500,00 0, Total Cost 478,19 319,61 305,90 324,28 365,56 411,09 459,27 509,80 561,87 614,64 Process Time 119, , , , , , , , , ,7269 Service Cost 58,11 115,99 174,42 230,65 290,00 348,92 407,46 465,97 523,88 579,00 VOT Cost 420,08 203,62 131,48 93,63 75,56 62,16 51,81 43,82 37,99 35,65 Total plus Delay Cost 3674,95 319,61 305,90 324,28 365,56 411,09 459,27 509,80 561,87 614,64 Figure 4: Number of Passengers

96 5000, , , , , ,00 DP , , ,00 500,00 0, Total Cost 538,96 359,00 339,36 361,18 402,82 445,90 497,88 551,24 610,28 665,66 Process Time 129, , , , , , , , , ,5568 Service Cost 62,95 126,47 190,14 252,76 317,63 377,91 441,43 503,62 568,83 627,53 VOT Cost 476,01 232,53 149,22 108,42 85,19 67,99 56,45 47,62 41,45 38,13 Total plus Delay Cost 4431,95 359,00 339,36 361,18 402,82 445,90 497,88 551,24 610,28 665,66 Figure 5: Number of Passengers , , ,00 DP , , ,00 0, Total Cost 598,97 392,28 369,05 392,91 435,63 481,95 534,11 597,75 650,77 715,58 Process Time 139, , , , , , , , , ,5666 Service Cost 67,92 135,80 203,67 271,55 340,85 405,92 471,03 543,59 605,82 674,87 VOT Cost 531,05 256,48 165,38 121,36 94,79 76,03 63,08 54,16 44,95 40,71 Total plus Delay Cost 5208,63 392,28 369,05 392,91 435,63 481,95 534,11 597,75 650,77 715,58 Figure 6: Number of Passengers

97 6000, , ,00 DP , , ,00 0, Total Cost 630,61 414,05 388,95 411,77 453,61 503,52 561,26 622,50 685,75 750,31 Process Time 144, , , , , , , , , ,0490 Service Cost 70,53 141,55 212,60 284,50 353,68 423,38 495,05 565,47 637,44 707,33 VOT Cost 560,09 272,50 176,35 127,26 99,92 80,13 66,21 57,03 48,31 42,98 Total plus Delay Cost 5614,53 414,05 388,95 411,77 453,61 503,52 561,26 622,50 685,75 750,31 Figure 7: Number of Passengers , , , , ,00 DP , , , ,00 0, Total Cost 852,09 548,71 504,49 528,29 577,66 636,20 712,97 786,31 861,01 946,22 Process Time 180, , , , , , , , , ,7019 Service Cost 88,80 177,88 265,74 356,00 443,27 530,82 621,16 709,60 796,76 890,62 VOT Cost 763,29 370,83 238,75 172,28 134,39 105,38 91,81 76,72 64,25 55,61 Total plus Delay Cost 8468, ,60 504,49 528,29 577,66 636,20 712,97 786,31 861,01 946,22 Figure 8: Number of Passengers

98 12000, , , ,00 DP , ,00 0, Total Cost 965,83 617,16 565,38 590,16 641,50 711,62 789,69 870,27 959, ,69 Process Time 199, , , , , , , , , ,3901 Service Cost 98,14 196,15 293,35 392,57 489,00 588,18 685,53 782,45 884,17 983,05 VOT Cost 867,69 421,01 272,03 197,59 152,49 123,44 104,16 87,82 75,52 63,64 Total plus Delay Cost 9927, ,17 565,38 590,16 641,50 711,62 789,69 870,27 959, ,69 Figure 9: Number of Passengers , , , ,00 DP , , ,00 0, Total Cost 1216,29 771,43 703,59 725,67 785,63 861,79 960, , , ,66 Process Time 240, , , , , , , , , ,0480 Service Cost 118,81 237,38 356,89 474,68 592,21 707,57 831,48 948, , ,55 VOT Cost 1097,48 534,05 346,70 250,99 193,42 154,23 129,34 108,79 91,28 78,11 Total plus Delay Cost 13152, , ,79 725,67 785,63 861,79 960, , , ,66 Figure 10: Number of Passengers

99 20000, , , , , ,00 DP , , , ,00 0, Total Cost 1678, ,60 939,46 957, , , , , , ,99 Process Time 309, , , , , , , , , ,7851 Service Cost 153,37 306,71 460,27 611,14 766,22 921, , , , ,71 VOT Cost 1525,30 740,89 479,19 346,54 268,21 213,54 173,70 146,26 124,90 109,28 Total plus Delay Cost 18593, , , , , , , , , ,99 Figure 11: Number of Passengers , , , , ,00 DP , , ,00 0, Total Cost 3164, , , , , , , , , ,31 Process Time 539, , , , , , , , , ,1164 Service Cost 268,10 536,57 804, , , , , , , ,71 VOT Cost 2896, ,39 913,97 663,08 509,64 411,01 334,41 282,02 237,49 197,60 Total plus Delay Cost 36601, , , , , , , , , ,31 Figure 12: Number of Passengers

100 APPENDIX E: Witness Software Outputs Statistics (140 Pass. 1 Check-in) Flight Type Domestic Passenger Number 140 Check-in Number 1 Figure 1: Passenger Flow 100

101 Activity Statistics Report by Total Simulation Time Baggagecontrol %Idle %Cycle - Busy %Cycle - Filling %Cycle - Emptying %Stopped - Blocked %Waiting - Setup %Stopped - Down %Waiting - Cycle %Stopped - Setup %Waiting - Repair Off-Shift Figure 2: Baggage Control Statistics Figure 3: Check-in utilization (%) 101