Heuristic approaches for operational aircraft maintenance routing problem with maximum flying hours and man-power availability considerations
| Pages | 2142-2170 |
| DOI | https://doi.org/10.1108/IMDS-11-2016-0475 |
| Published date | 04 December 2017 |
| Date | 04 December 2017 |
| Author | Abdelrahman E.E. Eltoukhy,Felix T.S. Chan,S.H. Chung,Ben Niu,X.P. Wang |
| Subject Matter | Information & knowledge management,Information systems,Data management systems,Knowledge management,Knowledge sharing,Management science & operations,Supply chain management,Supply chain information systems,Logistics,Quality management/systems |
Heuristic approaches for
operational aircraft maintenance
routing problem with maximum
flying hours and man-power
availability considerations
Abdelrahman E.E. Eltoukhy, Felix T.S. Chan and S.H. Chung
Department of Industrial and Systems Engineering,
The Hong Kong Polytechnic University, Hung hom, Hong Kong
Ben Niu
College of Management, Shenzhen University, Shenzhen, China, and
X.P. Wang
Dalian University of Technology, Dalian, China
Abstract
Purpose –The purpose of this paper is twofold. First, to propose an operational model for aircraft
maintenance routing problem (AMRP) rather than tactical models that are commonly used in the literature.
Second, to develop a fast and responsive solution method in order to cope with the frequent changes
experienced in the airline industry.
Design/methodology/approach –Two important operational considerationswereconsidered,
simultaneously. First one is the maximum flying hours, and second one is the man-power availability. On
the other hand, ant colony optimization (ACO), simulated annealing (SA), and genetic algorithm (GA)
approaches were proposed to solve the model, and the upper bound was calculated to be the criteria to assess
the performance of each meta-heuristic. After attempting to solve the model by these meta-heuristics, the
authors noticed further improvement chances in terms of solution quality and computational time. Therefore,
a new solution algorithm was proposed, and its performance was validated based on 12 real data from the
EgyptAir carrier. Also, the model and experiments were extended to test the effect of the operational
considerations on the profit.
Findings –The computational results showed that the proposed solution algorithm outperforms other meta-
heuristics in finding a better solution in much less time, whereas the operational considerations improve the
profitability of the existing model.
Research limitations/implications –The authors focused on some opera tional considerations ra ther
than tactical conside rations that are commo nly used in the literatur e. One advantage of this i s that it
improves the profitabi lity of the existing mode ls. On the other hand, iden tifying future resea rch
opportunities shoul d help academic researchers to deve lop new models and improve the perf ormance of the
existing models.
Practical implications –The experiment results showed that the proposed model and solution methods are
scalable and can thus be adopted by the airline industry at large.
Originality/value –In the literature, AMRP m odels were cast with approximated assum ption regarding
the maintenance issu e, while neglecting the man-power avai lability consideration. However, in t his paper,
the authors attempted t o relax that maintenanc e assumption, and consi der the man-power avail ability
constraints. Since the result showed that these considerations improve the profitability by
5.63 percent in the larges t case. The proposed operat ional considerations a re hence significant. Als o,
the authors utilized A CO, SA, and GA to solve the model for the first time , and developed a new solution
algorithm. The value and signi ficance of the new algorithm app eared as follow. First, the soluti on
quality was improved si nce the average improvem ent ratio over ACO, SA, and G A goes up to 8.30,
Industrial Management & Data
Systems
Vol. 117 No. 10, 2017
pp. 2142-2170
© Emerald PublishingLimited
0263-5577
DOI 10.1108/IMDS-11-2016-0475
Received 4 November 2016
Revised 13 January 2017
Accepted 2 February 2017
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm
The work described in this paper was supported by The Natural Science Foundation of China
(Grant No. 71471158, 71571120, 71271140), grants from The Hong Kong Polytechnic University
(Project Nos. G-UB03, G-UA4F, G-YBFD, G-YBN1), and a grant from the Research Committee of
The Hong Kong Polytechnic University under student account code RTYN.
2142
IMDS
117,10
4.45, and 4.00 percent, r espectively. Second , the computational ti me was significantly impr oved since it
does not go beyond 3 seconds in all the 12 rea l cases, which is considered muc h lesser compared to
ACO, SA, and GA.
Keywords Simulated annealing, Genetic algorithm, Ant colony optimization,
Aircraft maintenance routing problem
Paper type Research paper
Nomenclature
Sets
NF the set of flight legs, indexed by i,j.
Kthe set of aircraft, indexed by k.
MT the set of maintenance stations,
indexed by m.
NFM the set of flight legs that their
destination offers a maintenance
check.
Athe set of airports, indexed by a.
othe dummy source node of the
network.
tthe dummy sink node of the
network.
Parameters
DT
i
the departure time of flight leg i.
AT
i
the arrival time of flight leg i.
TRT the turn-around time, which is
consumed for getting passenger
off, unloading the luggage,
changing the gate, boarding,
loading the luggage, and fuel the
aircraft.
O
ia
the origin binary indicator of flight
leg isuch that O
ia
¼1 if the origin
of flight leg iand the airport aare
the same, and 0 otherwise.
D
ia
the destination binary indicator of
flight leg isuch that D
ia
¼1 if the
destination of flight leg iand the
airport aare the same, and 0
otherwise.
FT
i
the flight duration of flight leg i.
v
ij
the through value of the
connection between flight legs i
and j.
T
max
the maximum flying hours
between two consecutive
maintenance checks.
MP
m
the man-power group that available
in maintenance station m.
ET
m
the close time for the maintenance
station m.
MT the time required to perform the
maintenance check.
RTAM
k
the ready time for aircraft kto
cover the next flight leg after
finishing the maintenance check.
Usually, it is equal the starting
time for maintenance plu s the
maintenance time.
Ma considerable big number.
MC
k
the maintenance cost paid for
maintaining aircraft k.
α
k
the aircraft delay binary indicator
such that α
k
¼1, if the aircraft has
a delay in the maintenance station
because of unavailability of the
man-power groups, and 0
otherwise.
PC
k
the penalty cost paid if the aircraft
khas a time delay in the
maintenance station.
Decision variables
xijk ¼
1 if the flight legs iand jare
covered by aircraft k
0 otherwise
8
>
<
>
:
yijk ¼
1 if aircraft kcovers flight legs iand j;
then undergo maintenance
0 otherwise
8
>
<
>
:
1. Introduction
The aviation industry is now faced with the challenge of achieving a highprofit margin with
the incessant increase in oil price, labor, and capital. In this regard, Liang and
Chaovalitwongse (2012) stated that US passenger airlines lost around $3.4 billion in 2009.
Such unpleasant situations have propelled airline companies toward solving the scheduling
2143
Heuristic
approaches for
operational
AMRP
To continue reading
Request your trial