Crew scheduling is one of the most common tasks in air transportation. There is a list of specific requirements for crew planning. Crew should be balanced between load and rest periods during flight duty period and duration of recovery time. Human action is limited by its performance. Increased load of a person without a required period of recovery will increase multiple risks associated with human factor in civil aviation. Therefore, after each flight which is associated with flight duty period, crew have to spend some time for recovery (which includes acclimatization and rest periods). International aviation community developed a list of specific requirements for the duration of flight duty period and required time for crew recovery. Airlines have to follow all these requirements to provide safe air transportation services. In the paper, we propose an algorithm for an automatic crew scheduling system based on effective time distribution. A history of flight duty periods assigned to a particular crew member has been accumulated and used as input data. The algorithm provides effective crew planning of airlines based on normative regulations provided in different countries. Specific software has been developed to verify proposed model of airline crew planning tasks.
Human factor action could be controlled by specific rules based on human responsibility in the air transportation system [15, 16]. To perform main tasks crew should have normal human needs for sleep and rest. Commonly, airlines have a special department responsible for organizing crews for a flight ( Crew Planning . Crew planning grounds on multiple normative documents which minimized human factor action and ensured required level of flight safety [17, 18]. Minimization of human factor action on the crew is achieved by multiple inspections and training [19, 20].
The crew selection process requires processing a large volume of data, including full flight history and results of last training by each personnel (pilot, flight attendant, engineer) [21, 22]. Crew selection process has to identify the most effective crew composition, which could be the most suitable based on the conditions of the specific flight under consideration and the flight schedule as an operational scale, from today to a week ahead (or in strategic from a week to a year). Selection process requires exchange of official data between a wide range of departments and includes a lot of nuances and bureaucracy for the crew planning manager [23, 24]. Modern information technologies have greatly facilitated this issue due to the high performance of storing data from various departments in one unified database and then using software to reproduce and represent data in a given format for one or another user [25].
Also, there are several specific software, that could automatically analyze changes to the flight schedule made by the flight planning manager of the flight planning department, the airline's current boards and crews [26, 27]. Result of data processing could be effective crew scheduling to serve particular flights. Also, such software supports simultaneous data sharing with all users, which has been involved in the process.
Specific software makes possible to specify different types of actions that can be assigned to the crew. For example, a train transfer or a scheduled air safety briefing is convenient, but the downside is that even though there is a function that shows the crew suitable for the selected flight, only takes into account the fact that they do not have other tasks during the selected flight. In the case if user puts such an action as a transfer for a certain crew member, the function will select him for the flight, although according to the rules, he cannot perform it due to a violation of the pre-flight rest.
In this paper, we develop an algorithm of passive decision support systems for implementation in specific crew planning software. Proposed algorithm uses input data from airline historical flight database to select crew members who are suitable for the flight according to multiple criteria and provides effective decisions based on the ranking system.
The paper is structured as follows: the second section includes detailed information about the organization and selection of the crew for the corresponding flight, as well as specific of flight planning department; the third section describes a mathematical and functional model of software, as well as the list of criteria for initial selection and ranking; fourth section escribe results regarding performance and quality of proposed software.
Air transportation grounds on aviation law and regulation of all activities, like any other business or human activity. Specific of air transportation grounds on airspace usage. Airspace is provided by different administrative bodies to airspace users based on criteria of safety of air transportation. Most flights use multiple airspaces to perform flight connections [28, 29].
Basic principles of air transportation and flight rules were approved at "Chicago Convention". The International Civil Aviation Organization (ICAO) regulates all activities of civil aviation [30]. ICAO issues Standards and Recommended Practices (SARPs) in the field of civil aviation, based on which ICAO members provide legislation regulating all aviation activities. SARPs are not mandatory and any country can notify a difference in its legislation and SARPs. However, since the uniformity of the application of these standards improves the state of aviation safety, more and more countries apply them. In addition, various international and regional organizations (Eurocontrol, IATA, and EASA) also issue rules for their participants.
In this study, we follow ICAO regulation on fatigue risk management, system requirements [17], and the European Union document on the regulation of air operations [31]. Normative documents define specific terms: Duty, Duty period, Flight duty period, and On-call to identify actions and duties of crew. When the flight planning manager assigns a crew member to a flight on a certain day, he assigns him a Duty period, within which he will perform the duties assigned to him. At the same time, the Duty period (DP) is not a Flight duty period (FDP), but the second may be included in the first. It could be, that when a crew member leaves home and expects to be taken to the airport, his Duty period continues, and when the airplane on which crew member takes off, his Flight duty period begins and ends when the plane lands and stops the engines. But his duty period ends when he leaves the airport and is taken home.
In addition, a crew member can be assigned On-call or standby as it is called otherwise, that is, he does not fly on that day, but at the same time remains in full readiness to arrive at the designated place and start performing duty. Of course, it is impossible to work 24/7 without a break, so there is a rest period when a crew member rests and cannot be called to perform duty.
Duration of all of these periods is described in normative documents. The maximum duty period is 14 hours per day, but the number of hours can be up to 16 with a reinforced crew and up to 18 with a doubled crew. Also, the maximum number of hours should not exceed 60 for 7 days, 110 for 14 days, and 190 for 28 days. The daily maximum flight duty period is determined based on time start, whether the crew has acclimatized, and the duration of the planned flight (the crew can perform several flights during the Flight duty period if it meets the standards) in accordance with Table 1 [19]. The number of sectors indicates the number of joint flights for one crew team.
Maximum flight duty time is 100 hours for 28 calendar days, 900 hours for 1 calendar year, and 1000 hours for 12 consecutive calendar months.
Acclimatization is another important element of human factor. Human require some time to adapt to new time zone. In case, after the flight, a crew member is not at the airport from where his flight started, the difference DTZ in the time zones between the point of departure and landing has to be calculated for getting a particular time for rest and recovery: ≤ 4, then crew requires from 48 to 72 hours for acclimatization; ≤ 6, then from 72 to 96 hours of rest are required; ≤ 9, then from 96 to 120 hours of rest could be required; ≤ 12, then 120 hours of rest for acclimatization could be provided.
If a crew member was not given enough hours of rest before the next flight duty period, crew should be considered in an uncertain state of acclimatization. That could be a reason for the increased risk of human factor action.
Duration of the rest period should be not less than the previous flight duty period or 12 hours, whichever is longer (in case if crew is still at the departure airport). There are also recurrent extended recovery rest periods, which are at least 36 hours of rest for 7 consecutive calendar days, while this time must include 2 local nights, and no more than 168 hours must pass from the end of one recurrent extended recovery rest period to the beginning of another. There are also mandatory pre-flight and post-flight 12-hour rest periods when the crew cannot perform additional flights.
Crew selection is an important element of risk minimization in civil aviation. Each crew member has a specific identification code. All historical data associated with the crew are archived in the specific database. Selection process should search for the most efficient set of crew members that meet the requirements of DP, FDP, rest, and acclimatization.
Input data for selection algorithm includes airplane crew schedules during particular periods which are provided by specific software. Proposed algorithm analysis of input data and provides a ranked list of crew for upcoming flights for a specified duration of flight planning. User (or flight planning manager) of software should choose a crew load for the upcoming period. Following the selected parameters, the input array of data is sorted.
Let's denote the input associative array of data as A, which consists of keys that represent unique records of the number of crew members ai. Value is assigned in the form of tuples of data Ni concerning the tasks assigned to particular crew members for a selected period.
The structural scheme of proposed algorithm is shown in Figure 1. Input data obtained from the database includes crew members' identification and their schedules. Also, input data includes information from flight planning manager about the list of events that must be taken into account when calculating the FDP. Selection is an iterative process of removing the worst crew which is continued until the required crew set is obtained.
Each task has a specific duration which includes the datetime of its begin and finish. After initial screening of input data by identification code of crew member, a ranked list of crew members could be obtained. That restriction for FDP, rest, and acclimatization should be checked for each line of the sorted list. In case, if requirements are met and the selected crew member can perform the flight, he remains in the array. If not, crew member is removed from the array. Flight planning manager could assign other actions which could be other than boarding the flight, but these actions do not affect the increase of the corresponding employee's flight time. It is also possible to assign the following types of actions to particular array D and then calculate for each person his raid (FDPi) for the selected period:
= ∑ =1 ( , ), ( , ) = { 0, ∃ . ∄
,
= 1 − 2 ,
(
Value of FDP is calculated by (
Database with flights and crew
members
Crew members date Sorting of crew members by admission to the selected type of aircraft
Calculation of FDPi Selection according to the laws regarding FDP and on days off
Assessment of FDP potential in 28 days reating a ranked list
Data on the selected flight A set of event
types to calculate, FDPi International aviation norms
and laws
A set of event types to calculate,
Pi
Flight plannin
g manager 0, ∃ . ∄ , (4) (5) where is the ratio of Flight duty of the i-th crew member to the difference between all the time in 28 calendar days and the time when he could not perform Flight duty due to other events.
Finally, value could be used as a weight coefficient for crew members' ranking. Based on the obtained results flight planning manager could choose an appropriate crew team for a particular flight. Complete algorithm of crew scheduling could be represented in form of pseudo code as a sequence of following steps: import DB, D, B, Norms // config parameter of connection to DB, diction of norms for FDP , rest and DP, A set of event types to calculate FDPi and Pi flight=input('Flight parameter:' // input flight planning manager information of flight 'Time of start:' 'Time of end:' 'Airport of start:' 'Airport of end:' 'Type of aircraft:') A=connect(DB).request(f'Select crew From BD where type_aircraft={flight["type"]} and airport of start={flight["start_airport"]} '
f'and time between "{flight["time_start"]-datetime.timedelta(days=28)}" and "{flight["time_start"]}";') // request to the database to obtain suitable crew members _A_=dict() // initialization of the dictionary where the selected crew members will be recorded for a in A: // selection of the list of all crew members
FDPi, P_pre,Rest=0,0,False // initializing changes for the current crew member for Ti in a: // list of tasks that were assigned to the crew member a if Ti.type not in D:
FDPi+=Ti.finish-Ti.start // FDPi calculation if Ti.type not in B:
P_pre +=Ti.finish-Ti.start // calculation of how many employees were unavailable this month if flight["time_start"]-Ti.start<datetime.timedelta(days=7): if Ti.type =='Rest':
F_rest(Ti,Rest) // Rest calculation for the last 7 days before the flight Pi=FDPi/(28*24-P_pre) // calculation of Pi if FDPi +(flight['time_end']-flight['time_start'])>=Norms['FDP'] or not Rest: // Verification of compliance with norms
_A_.update({a:[Pi,FDPi]}) // adding the selected user to the dictionary _A_.sort() // sorting the dictionary from smallest to largest value print(_A_) // flight planning manager value output
Based on proposed algorithm a specific software has been developed in Python. MySQL server is used for data base.
Validation of proposed algorithm has been done with developed software in Python. Also, we use real flight planning data of a particular airline. We consider airline staff of 80 persons, 20 of them will be assigned to 6 upcoming flights based on their schedule data for the past 28 days before the date of the selected flights. We consider two cases of the flight planning process: manual mode when flight planning manager analyses all the data manually to prepare a crew list and automatically with proposed software.
The time required for the selection of a crew list has been measured for different numbers of flights (each flight requires a crew list). Performance of developed software in comparison to human performance is given in Figure 2.
The approximate processing time with 80 available crew members per flight, assuming the flight planning manager is not familiar with the situation per day, would be 10 minutes for manual mode. In addition, manual mode requires a lot of paper work, which can significantly increase the processing time. However, when the flight planning manager familiarizes himself with the flight situation on the day of the appointment, as well as reducing the crew list during selection, because if the first flight was selected from 80, the last from 55, the selection time will be reduced. Therefore, the process of selecting crew members for 6 flights of 80 crew members in manual mode will take up to 47 minutes. This value was obtained experimentally as the average value of the task completion time of four different flight planning managers in Table 2.
Developed software required at least 1 min 30 s to process crew list for one flight, and for 6 flights performance reached 10 min. Time of computation is a result of number of cycles performed by software when screening unsuitable crew members and calculating their FDPi. Increase in the number of crew members, computation performance decreases significantly. Also, obtained result shows that performance of software selection process is 5 times better than the manual one. In total flight planning manager selects a crew for 1 day in almost 50 minutes and for a week in 350 minutes or almost 6 hours of work, respectively, for 28 days, it is 1400 minutes or 23 hours, and for 20 minutes, it is 4 eight-hour working days. Software requires only 4 hours and 40 minutes to perform the same task. Flight planning manager has only to control the selection process and verify obtained results. Correlation of obtained results from humans and software is given in Tables 3 and 4.
Based on the above data, crew selection process by software mostly corresponded to results indicated by the flight planning manager. Three times out of 6 flights results of crew selection list are identical. Analysis of selection logic helps to identify three cases when crew lists were different: Computation of 2 flight includes several flight attendants with the same Pi. In the case of software, results have been obtained based on position in the array. In the case of human selection process, a choice for crew in this case has been given randomly.
A fourth flight considers a case when the training of the flight attendant should be carried out and, accordingly, one member of the crew will be an instructor who will perform the check and the other will be a flight attendant who will pass the check.
In flight 5 a co-pilot will be involved in another flight for which he will have to make a transfer to another airport and, accordingly, he cannot be involved in flight 5.
Results of verification give an 88% correlation between software results and human activity. However, time saved by software is significant. Also, final validation of results should be provided by humans at the end.
Crew planning process is an important element of civil aviation safety. Following all requirements of normative documents is mandatory for each crew member. Modern computerized systems mostly focus on only flight duty period calculation, but not the overall duty period associated with involved crew members. Proposed algorithm takes into account both flight duty and duty periods for scheduling crew members. Results of practical evaluation of proposed algorithm in Python indicate at least five times speedy than manual flight crew scheduling. Software provides efficient crew scheduling, however, obtained results need to be validated by flight planning managers to avoid mis selection. In the case when provided by software crew rank is identical to another crew member, flight planning manager could take into account the personal properties of each crew member (or preference for teamwork) to make an effective crew set for the particular flight.
In case the number of crew is less than 50 manual mode is possible, however for airlines with much bigger staff automatic crew scheduling is an important tool. The main advantage is the ability to quickly calculate additional parameters such as duty period, and flight duty period, which, with a large number of flights, allows for removal from the crew list, persons who are close to exceeding the norm, and with a small number of flights.
Automated crew scheduling can improve the results of the crew planning department due to faster selection of eligible persons for the flights to be performed by the airline. Proposed algorithm could be useful for both strategic and operational crew planners. Advances in Civil Aviation Systems Development. Lecture Notes in Networks and Systems, Springer, Cham, 2024, vol. 992, pp. 17 30, doi:10.1007/978-3-031-60196-5_2. [2] Annual Review 2023, International Air Transport Association, 2023. [3] I.V. Ostroumov, O. Ivashchuk, N.S. Kuzmenko, Preliminary Estimation of war Impact in Ukraine on the Global Air Transportation, in: Proceedings of IEEE 12th International Conference on Advanced Computer Information Technologies (ACIT), Ruzomberok, Slovakia, 2022, pp. 281 284, doi: 10.1109/ACIT54803.2022.9913092. [4] Global fleet size analysis by ch-aviation, June 2022. URL: https://about.chaviation.com/blog/2022/06/30/june-2022-global-fleet-size-analysis-by-ch-aviation. [5] I.V. Ostroumov, V.P. Kharchenko, N.S. Kuzmenko, An airspace analysis according to area navigation requirements, Aviation 23 (2) (2019) 36 42, doi: 10.3846/aviation.2019.10302. [6] I. Ostroumov, N. Kuzmenko, Risk Assessment of Mid-air Collision Based on Positioning Performance by Navigational Aids, in: Proceedings of IEEE 6th International Conference on Methods and Systems of Navigation and Motion Control (MSNMC), Kyiv, Ukraine, 2020, pp. 34 37, doi: 10.1109/MSNMC50359.2020.9255506. [7] Y. Averyanova, V. Larin, N. Kuzmenko, M. Zaliskyi, O. Solomentsev, Turbulence Detection and Classification Algorithm Using Data from AWR, in: Proceedings of IEEE 2nd Ukrainian Microwave Week (UkrMW), Ukraine, 2022, pp. 518 522. doi: 10.1109/UkrMW58013.2022.10037172. [8] N. Kuzmenko, I. Ostroumov, Y. Bezkorovainyi, O. Sushchenko, Airplane Flight Phase Identification Using Maximum Posterior Probability Method, in: Proceedings of IEEE 3rd International Conference on System Analysis & Intelligent Computing (SAIC), Kyiv, Ukraine, 2022, pp. 1 5, doi: 10.1109/SAIC57818.2022.9922913. [9] I. V. Ostroumov, N. S. Kuzmenko, An Area Navigation (RNAV) System Performance Monitoring and Alerting in: Proceedings of IEEE First International Conference on System Analysis & Intelligent Computing (SAIC), Kyiv, UKraine, 2018, pp. 1 4. doi: 10.1109/SAIC.2018.8516750. [10] I.V. Ostroumov, N.S. Kuzmenko, Compatibility analysis of multi signal processing in APNT with current navigation infrastructure, Telecommunications and Radio Engineering, 77(3) (2018) 211 223. doi: 10.1615/TelecomRadEng.v77.i3.30. [11] Z. Lunlong, L. Jiongpo, Analysis of requirements and architecture of the next generation flight management system, in: Proceedings of IEEE 7th International Conference on Information Science and Control Engineering (ICISCE), Changsha, China, 2020, pp. 1678 1682. doi: 10.1109/ICISCE50968.2020.00332. [12] T. Kilbourne, F. Wieland, M. Lehky, B. Trainum, M. Underwood, Toward a Cloud-Based Flight Management System, in: Proceedings of IEEE/AIAA 40th Digital Avionics Systems Conference (DASC2021), San Antonio, TX, USA, 2021, pp. 1 8. doi: 10.1109/DASC52595.2021.9594466. [13] M. Zaliskyi, O. Solomentsev, V. Larin, Y. Averyanova, N. Kuzmenko, Model Building for Diagnostic Variables during Aviation Equipment Maintenance, in: Proceedings of IEEE 17th International Conference on Computer Sciences and Information Technologies (CSIT), Lviv, Ukraine, 2022, pp. 160 164, doi: 10.1109/CSIT56902.2022.10000556. [14] O. Solomentsev, M. Zaliskyi, O. Holubnychyi, O. Sushchenko, Y. Bezkorovainyi, Efficiency Analysis of Current Repair Procedures for Aviation Radio Equipment, in: I. Ostroumov, M. Zaliskyi (Eds.), Proceedings of the International Workshop on Advances in Civil Aviation Systems Development. Lecture Notes in Networks and Systems, Springer, Cham, 2024, vol. 992, pp. 281 295. doi: 10.1007/978-3-031-60196-5_21. [15] W. T. K. Chan, W. C. Li, Development of effective human factors interventions for aviation safety management. Frontiers in public health 11 (2023) 1144921. doi: 10.3389/fpubh.2023.1144921. [16] D. A. Sant, A. V. Hilal, The impact of human factors on pilots safety behavior in offshore aviation companies: A Brazilian case. Safety science 140 (2021) 105272. doi: 10.1016/j.ssci.2021.105272. [17] Manual for the Oversight of Fatigue Management Approaches, Doc. 9966, ICAO, 2016. [18] Fatigue Management Guide for Airline Operators, IATA, 2015. [19] Operation of Aircraft, Annex 6, Standards and Recommended, ICAO, 2018. [20] O. Holubnychyi, M. Zaliskyi, O. Sushchenko, O. Solomentsev, Y. Averyanova, Self-Organization Technique with a Norm Transformation Based Filtering for Sustainable Infocommunications Within CNS/ATM Systems, in: I. Ostroumov, M. Zaliskyi (Eds.), Proceedings of the International Workshop on Advances in Civil Aviation Systems Development. Lecture Notes in Networks and Systems, Springer, Cham, 2024, vol. 992, pp. 262 278. doi: 10.1007/978-3-031-60196-5_20. [21] X. Wen, X. Sun, Y. Sun, X. Yue, Airline crew scheduling: Models and algorithms. Transportation research part E: logistics and transportation review 149 (2021) 102304. doi: 10.1016/j.tre.2021.102304. [22] X. Wen, S. H. Chung, H. L. Ma, W. A. Khan, Airline crew scheduling with sustainability enhancement by data analytics under circular economy. Annals of Operations Research (2023) 1 27. doi: 10.1007/s10479-023-05312-7. [23] Y. Xu, S. Wandelt, X. Sun, Airline scheduling optimization: literature review and a discussion of modelling methodologies. Intelligent Transportation Infrastructure, 3 (2024) liad026. doi: 10.1093/iti/liad026. [24] X. Wen, H. L. Ma, S. H. Chung, W. A. Khan, Robust airline crew scheduling with flight flying time variability. Transportation Research Part E: Logistics and Transportation Review 144 (2020) 102132. doi: 10.1016/j.tre.2020.102132. [25] O. Solomentsev, M. Zaliskyi, O. Sushchenko, Y. Bezkorovainyi, Y. Averyanova, Data Processing through the Lifecycle of Aviation Radio Equipment, in: Proceedings of IEEE IEEE 17th International Conference on Computer Sciences and Information Technologies (CSIT), 2022, pp. 146 151. doi: 10.1109/CSIT56902.2022.10000844. [26] X. Sun, S. H. Chung, H. L. Ma, Operational risk in airline crew scheduling: do features of flight delays matter? Decision Sciences 51(6) (2020) 1455 1489. doi: 10.1111/deci.12426. [27] W. Ouyang, X. Zhu, Meta-heuristic solver with parallel genetic algorithm framework in airline crew scheduling. Sustainability 15(2) (2023) 1506. doi: 10.3390/su15021506. [28] M. Zaliskyi, O. Solomentsev, O. Holubnychyi, I. Ostroumov, O. Sushchenko, Yu. Averyanova, Y.
Bezkorovainyi, K. Cherednichenko, O. Sokolova, V. Ivannikova, R. Voliansky, B. Kuznetsov, I. Bovdui, T. Nikitina, Methodology for substantiating the infrastructure of aviation radio equipment repair centers, CEUR Workshop Proceedings 3732 (2024) 136 148. https://ceur-ws.org/Vol-3732/paper11.pdf [29] Z. W. Zhong, Overview of recent developments in modelling and simulations for analyses of airspace structures and traffic flows. Advances in Mechanical Engineering, 10(2) (2018) 1687814017753911. doi: 10.1177/16878140177539. [30] Convention on International Civil Aviation, Doc 7300, 1944. [31] Technical requirements and administrative procedures related to air operations pursuant to Regulation (EC), No 216/2008, European Parliament and of the Council, 2014.