Hindawi Publishing Corporation

Journal of Applied Mathematics

Volume 2013, Article ID 742653, 8 pages
http://dx.doi.org/10.1155/2013/742653

Research Article
A Hybrid Metaheuristic for Multiple Runways Aircraft Landing
Problem Based on Bat Algorithm

Jian Xie,1 Yongquan Zhou,1,2 and Hongqing Zheng1

1
College of Information Science and Engineering, Guangxi University for Nationalities, Nanning, Guangxi 530006, China
2
Guangxi Key Laboratory of Hybrid Computation and IC Design Analysis, Nanning, Guangxi 530006, China

Correspondence should be addressed to Yongquan Zhou; [email protected]

Received 17 May 2013; Accepted 11 July 2013

Academic Editor: Xin-She Yang

Copyright © 2013 Jian Xie et al. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The aircraft landing problem (ALP) is an NP-hard problem; the aim of ALP is to minimize the total cost of landing deviation from
predefined target time under the condition of safe landing. In this paper, the multiple runways case of the static ALP is considered
and a hybrid metaheuristic based on bat algorithm is presented to solve it. Moreover, four types of landing time assignment
strategies are applied to allocate the scheduling time, and a constructed initialization is used to speed up the convergence rate.
The computational results show that the proposed algorithm can obtain the high-quality and comparable solutions for instances
up to 500 aircrafts, and also it is capable of finding the optimal solutions for many instances in a short time.

1. Introduction Metaheuristic includes genetic algorithms [7], ant colony

optimization [8], simulated annealing [9], scatter search and
Airport runway scheduling optimization is an ongoing bionomic algorithms [2], cellular automata optimization [10],
challenge for air traffic controllers. The increasing demand and other hybrid metaheuristics [9, 11]. Briskorn and Stolletz
is a challenge, which leads to many airports and routes resent integer programming models an aircraft landing prob-
are congested. Although investment in infrastructure may lems with aircraft classes [12].
increase capacity at airports, it is an effective solution to Recently, more and more metaheuristics inspired by
improve planning and scheduling on current infrastructure. nature or social phenomenon are proposed and these algo-
The aircraft landing Problem (ALP) is a kind of typical NP- rithms are increasingly applied to different fields. The bat
hard problem in airport runway scheduling optimization [1]. algorithm (BA) is one of the most popular algorithms, which
The ALP consists of determining an optimal schedule of is inspired by the intelligent echolocation behavior of micro-
landing aircrafts on runways and assigning the landing time bats when they are foraging [13]. Many researchers applied
of each arriving aircraft. The objective is to minimize the total BA to solve various optimization problems. For example,
cost of landing deviation from predefined target time under Gandomi et al. focus on solving constrained optimization
the condition of safe landing. A predefined time window and tasks [14]. Yang and Gandomi apply BA to solve many
separation time requirements with other aircraft must meet. global engineering optimizations [15]. Mishra at al. use BA
The first-come first-served (FCFS) is one of the most to update the weights of a functional link artificial neural
common solving methods for ALP, especially, in the sin- network classifier, a model proposed for classification [16].
gle runway situation. A detailed review of published work Meanwhile, some researchers have improved BA and applied
addressing the ALP can be found in [1–3]. Generally it to various optimization problems. Xie et al. proposed a
speaking, for ALP, those methods can be broadly classified bat algorithm based on differential operator and the Lévy
into three categories: exact methods (e.g., dynamic pro- flights trajectory (DLBA) to solve function optimization and
gramming [4], branch-and-bound [1], and branch-and-price nonlinear equations [17]. Wang et al. proposed a new bat
[5]), queueing theory [6], and heuristic or metaheuristic. algorithm with mutation (BAM) to solve the uninhabited
2 Journal of Applied Mathematics

combat air vehicle (UCAV) path planning problem [18]. In 𝜏𝑖⃗ = the delay of landing plane 𝑖
this paper, the multiple runways case of the static ALP is (landing after the target time) ;
considered; a hybrid metaheuristic based on bat algorithm
(HBA, for short) is presented to solve it. thus 𝜏𝑖⃗ = max (0, 𝑡𝑖 − 𝑇𝑖 ) ;
The rest of this paper is organized as follows. Section 2 ←
󳨀
𝜏 𝑖 = the earliness of landing plane 𝑖
presents the mathematical model of ALP and original bat
(landing before the target time) ;
algorithm. Hybrid metaheuristic based on bat algorithm is
proposed in detail to solve multiple runways aircraft landing thus ←𝜏󳨀 = max (0, 𝑇 − 𝑡 ) .
𝑖 𝑖 𝑖
problem in Section 3. The experimental results of the HBA (1)
and comparisons with other previous algorithms are shown
in Section 4. In the last section, we conclude this paper and The mathematical formulation of a mixed-integer pro-
point out some future work. gramming for this problem is as follows [1, 2, 9]:
𝑛
2. Problem Definitions and Bat Algorithm min Z = ∑ (𝜏𝑖⃗ 𝑐𝑖⃗ + ←
𝜏󳨀𝑖 ←
󳨀
𝑐 𝑖) (2)
𝑖=1
2.1. The Mathematical Formulation of ALP. The ALP aims
at finding the best arrangement of sequences, runway, and s.t. 𝐸𝑖 ≤ 𝑡𝑖 ≤ 𝑇𝑖 , 𝑖 = 1, 2, . . . 𝑛; (3)
corresponding landing time for a given set of landing aircraft
to minimize total cost by following separation requirements. 𝛾𝑖𝑗 + 𝛾𝑗𝑖 = 1, ∀𝑖, 𝑗 ∈ {1, 2, . . . , 𝑛} , 𝑖 ≠ 𝑗; (4)
To illustrate the mathematical formulation, some notations
and decision variables are defined as follows.
Notations: 𝑡𝑗 ≥ 𝑡𝑖 + 𝑆𝑖𝑗 𝛿𝑖𝑗 + 𝑠𝑖𝑗 (1 − 𝛿𝑖𝑗 ) − 𝑀𝛿𝑖𝑗 ;

𝑛 = the number of planes; 𝑀 = (𝐿 𝑖 + max (𝑆𝑖𝑗 , 𝑠𝑖𝑗 ) − 𝐸𝑗 ) , ∀𝑖 ≠ 𝑗 ∈ {1, 2, . . . , 𝑛} ;

𝑚 = the number of runways; (5)
𝑆𝑖𝑗 = the separation time (≥0) between plane 𝑖 landing 𝑚
and plane 𝑗 landing, (where the planes 𝑖 and 𝑗 land on ∑𝛾𝑖𝑟 = 1, ∀𝑖 ∈ {1, 2, . . . , 𝑛} ; (6)
the same runways), 𝑖 ≠ 𝑗 ∈ {1, 2, . . . , 𝑛}; 𝑟=1

𝑠𝑖𝑗 = the separation time (0 ≤ 𝑠𝑖𝑗 ≤ 𝑆𝑖𝑗 ) between plane 𝛿𝑖𝑗 ≥ 𝛾𝑖𝑟 +𝛾𝑗𝑟 −1, 𝑖 ≠ 𝑗 ∈ {1, 2, . . . , 𝑛} , 𝑟 ∈ {1, 2, . . . , 𝑚} ;
𝑖 landing and plane 𝑗 landing, (where the planes 𝑖 and (7)
𝑗 land on the different runways), 𝑖 ≠ 𝑗 ∈ {1, 2, . . . , 𝑛};
𝑇𝑖 = the target landing time (target time) of plane 𝑖, 𝑡𝑖 − 𝑇𝑖 = 𝜏𝑖⃗ − ←
𝜏󳨀𝑖 , 𝑖 = 1, 2, . . . , 𝑛;
𝑖 ∈ {1, 2, . . . , 𝑛}; 0 ≤ 𝜏𝑖⃗ ≤ 𝑇𝑖 − 𝐸𝑖 , 𝜏𝑖⃗ ≥ 𝑇𝑖 − 𝑡𝑖 , 𝑖 = 1, 2, . . . , 𝑛; (8)
𝐸𝑖 = the earliest landing time of plane 𝑖, 𝑖 ∈
{1, 2, . . . , 𝑛}; 0≤←
𝜏󳨀𝑖 ≤ 𝐿 𝑖 − 𝑇𝑖 , ←
𝜏󳨀𝑖 ≥ 𝑡𝑖 − 𝑇, 𝑖 = 1, 2, . . . , 𝑛;
𝐿 𝑖 = the latest landing time of plane 𝑖, 𝑖 ∈ {1, 2, . . . , 𝑛};
𝑐𝑖⃗ = the cost of late landing of plane 𝑖, 𝑖 ∈ {1, 2, . . . , 𝑛}; 𝑦𝑖𝑗 , 𝛾𝑖𝑟 , 𝛿𝑖𝑗 ∈ {0, 1} , 𝑖=𝑗̸ ∈ {1, 2, . . . , 𝑛} , 𝑟 ∈ {1, 2, . . . , 𝑚} ;
←󳨀
𝑐 = the cost of early landing of plane 𝑖, 𝑖 ∈
𝑖 𝑡𝑖 , 𝜏𝑖⃗ , ←
𝜏󳨀𝑖 ≥ 0, 𝑖 = 1, 2, . . . 𝑛.
{1, 2, . . . , 𝑛}. (9)
Decision variables: Objective function (2) minimizes the total cost of landing
𝑡𝑖 = the scheduled landing time of plane 𝑖, 𝑖 ∈ {1, 2, . . . , 𝑛} ; deviation from target time. The constraints (3) ensure that
the scheduled landing time of each aircraft lies within its
{ 1, if plane 𝑖 lands before time window; constraints (4) consider the landing order;
{ either aircraft 𝑖 or 𝑗 must land first; separation constraints
𝑦𝑖𝑗 = { plane 𝑗 𝑖 ≠ 𝑗 ∈ {1, 2, . . . , 𝑛} ;
{ must be ensured by constraints (5), where the role of 𝑀 is
{0, otherwise, to ensure that the equation is redundant if 𝑗 lands before
𝑖; the constraints (6) ensure that each aircraft should land
{ 1, if plane 𝑖 lands on only one runway; when aircrafts 𝑖 and 𝑗 are assigned to
{
𝛾𝑖𝑟 = { on runway 𝑟 𝑖 ∈ {1, . . . , 𝑛} , 𝑟 ∈ {1, . . . , 𝑚} ; land on the same runway, constraints (7) ensure that the
{
{ 0, otherwise, runways assigned to aircrafts 𝑖 and 𝑗 are identical; additional
constraints (8) are introduced in order to link 𝜏𝑖⃗ and ← 𝜏󳨀𝑖 to the
{ 1, if planes 𝑖 and 𝑗 decision variable 𝑡𝑖 ; constraints (9) ensure that the decision
{
𝛿𝑖𝑗 = { land on the same runway 𝑖 ≠ 𝑗 ∈ {1, 2, . . . , 𝑛} ; variables 𝑦𝑖𝑗 , 𝛾𝑖𝑟 , and 𝛿𝑖𝑗 are binary and the decision variables
{
{0, otherwise, 𝑡𝑖 , 𝜏𝑖⃗ , and ←
𝜏󳨀𝑖 are nonnegative, respectively.
Journal of Applied Mathematics 3

2.2. Bat Algorithm. The basic bat algorithm (BA) is a meta- Aircraft: 1 2 3 4 5 6 7 8 9 10
heuristic proposed by Yang in 2010 [13]. Under several ideal Runway: 1 3 2 2 3 2 1 2 1 3
rules, the BA has the following steps.
Figure 1: The representation of the candidate solutions.
Step 1. Initialize the bat population and other parameters, and
these initial individuals are evaluated.

Step 2. Each bat individual randomly selects a certain fre- of the candidate solutions for this particular problem. Each
quency 𝑓𝑖 of sonic pulse, and the position of bat individual is individual is a sequence 𝑆 (𝑆 = (𝑠1 , 𝑠2 , . . . , 𝑠𝑛 )) with integer
updated according to their selected frequency. The formulas number 𝑠𝑖 (𝑠𝑖 ∈ {1, . . . , 𝑚}), where the integer number
are as follows: represents the runway and the length of this sequence 𝑆 is the
𝑓𝑖 = 𝑓min + (𝑓max − 𝑓min ) 𝛽, number of aircrafts. For example, if we have three runways
and ten aircrafts, the coded individual with integer is 1 →
V𝑖𝑡 = V𝑖𝑡−1 + (𝑥𝑖𝑡 − 𝑥∗ ) 𝑓𝑖 , (10) 3 → 2 → 2 → 3 → 2 → 1 → 2 → 1 → 3 Figure 1 shows that the
aircrafts 1, 7, and 9 land on runway number 1, the aircrafts
𝑥𝑖𝑡 = 𝑥𝑖𝑡−1 + V𝑖𝑡 , 3, 4, 6 and 8 land on runway number 2, the aircrafts 2, 5, and
10 land on runway number 3.
where 𝑥𝑖𝑡 and V𝑖𝑡 represent the positions and velocities of
individual 𝑖 in a D-dimensional search space at generation 3.2. Landing Time Assignment. The assigned landing time
𝑡, 𝛽 ∈ [0, 1] is a random vector drawn from a uniform (ALT) of each aircraft is the very important step in the
distribution, and 𝑥∗ is the current global best location ALP; the purpose is to reduce the total cost of penalty
(solution) which is located after comparing all the solutions caused by all aircrafts. In this paper, the landing time is
among all the 𝑛 bats. Meanwhile, these new individuals are assigned based on the target landing time 𝑇𝑖 assigned to
evaluated. each aircraft. There are four types of assignment strategies
carried out in the proposed algorithm: forward assignment
Step 3. If a random number is greater than its pulse emission
strategy (FAS), backward assignment strategy (BAS), random
rate 𝑅, then a new position is generated around the current
forward assignment strategy (RFAS), and random backward
global best position for each individual, which is the equal
assignment strategy (RBAS). For each kind of assignment
to local search. The updating formula of local search adopts
strategy, firstly, all aircrafts are found out on each runway,
𝑥 = 𝑥∗ + 𝜀 × 𝐿 𝑡 , where 𝜀 ∈ [−1, 1] is a random number and
and the target landing time 𝑇 of these aircrafts is sorted in
𝐿 𝑡 = ⟨𝐿𝑡𝑖 ⟩ is the average loudness of all the bats at current
ascending order, namely, 𝑇𝑖1 ≤ 𝑇𝑖2 ≤ ⋅ ⋅ ⋅ ≤ 𝑇𝑖𝑝−1 ≤ 𝑇𝑖𝑝 , where
generation 𝑡.
𝑝 is the number of aircrafts on this runway.
Step 4. If the local search is effective and its loudness 𝐿 is
greater than a random number, then the new position is (i) FAS: for each two sorted aircrafts 𝑖𝑘 and 𝑖𝑘+1 on a
accepted and its pulse emission rate 𝑅 and loudness 𝐿 are runway, the ALT is 𝑡𝑖⃗ 𝑘 and 𝑡𝑖⃗ 𝑘+1 , separately. If 𝑇𝑖𝑘+1 <
updated, where pulse emission rate is increased and loudness 𝑡𝑖⃗ 𝑘 + 𝑆𝑖𝑘 𝑖𝑘+1 ; then 𝑡𝑖⃗ 𝑘+1 = 𝑡𝑖⃗ 𝑘 + 𝑆𝑖𝑘 𝑖𝑘+1 , otherwise 𝑡𝑖⃗ 𝑘+1 =
is decreased. 𝐿 𝑖 and 𝑅𝑖 are updated by 𝐿𝑡+1
𝑖 = 𝛼 × 𝐿𝑡𝑖 , 𝑅𝑖𝑡+1 = 𝑇𝑖𝑘+1 .
0
𝑅𝑖 × [1 − exp(−𝛾 × 𝑡)], where 𝛼, 𝛾 are constants. (ii) BAS: for each two sorted aircrafts 𝑖𝑘 and 𝑖𝑘+1 on a
←
󳨀 ←
󳨀
Step 5. If the termination criterion is met, then the algorithm runway, the ALT is 𝑡 𝑖𝑘 and 𝑡 𝑖𝑘+1 , separately. If 𝑇𝑖𝑘 >
←
󳨀 ←
󳨀 ←
󳨀
stops; otherwise repeat algorithm (go to Step 2). 𝑡 𝑖𝑘+1 − 𝑆𝑖𝑘 𝑖𝑘+1 ; then 𝑡 𝑖𝑘 = 𝑡 𝑖𝑘+1 − 𝑆𝑖𝑘 𝑖𝑘+1 , otherwise
←
󳨀
In general, the bat algorithm has three procedures, posi- 𝑡 𝑖𝑘 = 𝑇𝑖𝑘 .
tion updating, local search, and decreasing the probability of (iii) RFAS: for each two sorted aircrafts 𝑖𝑘 and 𝑖𝑘+1 on
local search. For the details of BA refer to [13]. ⌣ ⌣ ⌣
a runway, the ALT is 𝑡 𝑖𝑘 and 𝑡 𝑖𝑘+1 , separately. 𝑡 𝑖𝑘 is
←
󳨀
3. Hybrid Metaheuristic Based on Bat rounded down to the nearest integer between 𝑡 𝑖𝑘 and
⌣ ⌣ ⌣ ⌣
Algorithm for ALP 𝑡𝑖⃗ 𝑘 . If 𝑡 𝑖𝑘+1 < 𝑡 𝑖𝑘 + 𝑆𝑖𝑘 𝑖𝑘+1 , then 𝑡 𝑖𝑘+1 = 𝑡 𝑖𝑘 + 𝑆𝑖𝑘 𝑖𝑘+1 .

Basic bat algorithm is a continuous optimization algorithm, (iv) BFAS: for each two sorted aircrafts 𝑖𝑘 and 𝑖𝑘+1 on
which is successfully applied to solve real optimization a runway, the ALT is 𝑡𝑖̆ 𝑘 and 𝑡𝑖̆ 𝑘+1 , separately. 𝑡𝑖̆ 𝑘 is
←󳨀
problem [15, 17]. However, the standard continuous encoding rounded down to the nearest integer between 𝑡 𝑖𝑘 and
scheme of BA cannot be used to solve ALP directly. Therefore, 𝑡𝑖⃗ 𝑘 . If 𝑡𝑖̆ 𝑘 < 𝑡𝑖̆ 𝑘+1 − 𝑆𝑖𝑘 𝑖𝑘+1 , then 𝑡𝑖̆ 𝑘 = 𝑡𝑖̆ 𝑘+1 − 𝑆𝑖𝑘 𝑖𝑘+1 .
in order to solve aircraft landing problem effectively, HBA is
proposed. The assigned landing time is effective just while the all
constraints (3)–(9) are satisfied, and the four kinds of total
3.1. Solution Representation in HBA. In order to apply BA cost are compared; the best total cost is used as the objective
to ALP, the first step is to devise a suitable representation function value.
4 Journal of Applied Mathematics

3.3. Initialization Construction. The initialization of a popu-

Before: 1 3 2 2 3 2 1 2 1 3
lation is usually performed by randomly selecting a runway
from the available runways for each aircraft. However, the
target landing time of aircraft is ordered, and the time window After: 1 3 2 3 3 2 1 2 1 3
must ensure considering the safety. So, a preprocessing is
performed to improve the performance of HBA in the Figure 2: An individual with ten aircrafts before and after position
initialization construction. updating.
In initialization, 𝑝𝑠 individuals are generated, where 𝑝𝑠 is
population size. For each individual, all target landing time
𝑇𝑖 (𝑖 = 1, 2, . . . , 𝑛) is sorted in an ascending order; namely, Before: 1 3 2 2 3 2 1 2 1 3
𝑇𝑖1 ≤ 𝑇𝑖2 ≤ ⋅ ⋅ ⋅ ≤ 𝑇𝑖𝑛−1 ≤ 𝑇𝑖𝑛 . The first aircraft 𝑖1 lands on a
runway randomly, for each two sorted aircrafts 𝑖𝑘 and 𝑖𝑘+1 , if After: 1 3 3 2 3 2 1 2 1 2
𝑇𝑖𝑘+1 < 𝑇𝑖1 + 𝑆𝑖𝑘 𝑖𝑘+1 , and then aircraft 𝑖𝑘+1 is allocated to other
runway, namely, aircrafts 𝑖𝑘 and 𝑖𝑘+1 on different runways. Figure 3: An individual with ten aircrafts before and after swap local
Otherwise, the aircraft 𝑖𝑘+1 is allocated to the runway that search.
aircraft 𝑖𝑘 lands on. The initialization repeats until the 𝑝𝑠
individuals are generated.
According to the experiments performed, this initializa- purpose of the local search is to enhance the solution
tion construction can develop an initial schedule of land- generated, and the operation is performed on the current
ing aircrafts which has a very good quality. However, to global best individual in bat algorithm. In this paper, two
derive near-optimal solutions, this initial schedule should be types of local search are presented: the swap and loop sub-
improved using the bat algorithm provided in Section 3.4. sequence inserting (LSI).
The purpose of swap is mildly mutating the current
3.4. Hybrid Bat Algorithm. In original bat algorithm frame- global best individual, so that an improved solution can be
work, the idea is that, firstly, the bat individual randomly found out around the current optimal solution. The swap is
selects a certain frequency of sonic pulse, and the position of illustrated in Figure 3, where the third and tenth components
bat individual is updated according to its selected frequency; are randomly chosen to be exchanged; note that the two
secondly, if a random number is greater than its pulse selected aircrafts land on different runways. Thus, the value of
emission rate 𝑅, then a new position is generated around the third component is changed from 2 to 3, while the value
the current global best position for each individual, which of the tenth component is switched from 3 to 2.
is equal to local search; at last, if the local search is effective The LSI is a variant of inserting operation; the purpose is
and its loudness 𝐿 is greater than a random number, then to mutate the current global best individual in a large extent,
the new position is accepted, and its pulse emission rate so that the diversity of population can be ensured. It can
𝑅 and loudness 𝐿 are updated, where pulse emission rate effectively avoid prematurity and greatly improve efficiency
is increased and loudness is decreased. In general, the bat of global search. The LSI is illustrated in Figure 4; a start
algorithm has three procedures: position updating, local point is randomly chosen (ninth component) and a random
search, and decreasing the probability of local search. length of subsequence is determined (5), so the sub-sequence
In this paper, the frequency 𝑓 is a runway, 𝑓 ∈ {1, . . . , 𝑚}. can be determined (1 → 3 → 1 → 3). A random insert point is
The position updating is different from continuous bat chosen in remainder sub-sequence (third, fifth components);
algorithm. The position updating is used to assign the landing then the selected sub-sequence is inserted into remainder
sequence, which is performed as follows: sub-sequence before insert point.
The local search systematically explores different neigh-
(i) select a frequency 𝑓 randomly, namely; select a borhood structures. The swap and LSI are preformed accord-
runway randomly; ing to the pulse emission rate 𝑅. In other words, if a
(ii) select an aircraft randomly on selected runway; then random number is greater than the 𝑅, the swap is performed;
assign this aircraft to other runway. otherwise, the LSI is performed.
The loudness 𝐿 𝑖 of bat individual 𝑖 determines the
There is an example used to illustrate the procedure in accepted probability a solution generated by local search
Figure 2. If the frequency 𝑓 = 2 and the second aircraft is in original bat algorithm. Meanwhile, it also dominates
selected, then, this aircraft is assigned to runway number 3. the updating of pulse emission rate 𝑅 and loudness 𝐿 in
For the local search part, this procedure is controlled by continuous bat algorithm. However, in this paper, a runway
pulse emission rate 𝑅. The 𝑅 is equivalent to the probability balance (RB) operation is performed according to the value
of performing local search, and the 𝑅 is updated by of loudness 𝐿 for each individual.
The intention of RB operation is to balance the load of
−1
5 𝑡max each runway. Firstly, the aircrafts are counted on each run-
𝑅 (𝑡) = (1 + exp (− × (𝑡 − ))) , (11)
𝑡max 2 way; an aircraft selected randomly on runway with maximum
aircrafts is assigned to a runway with minimum aircrafts.
where 𝑡 denotes the 𝑡th generation 𝑡max is the maximal If the amount of aircrafts on runway (maximum aircrafts
generation. The rate 𝑅 is similar to sigmoid function. The or minimum aircrafts) is equal, then a runway is selected
Journal of Applied Mathematics 5

4.1. Sensitivity Analysis. Since performance is affected by

Before: 1 3 2 2 3 2 1 2 1 3
the settings of the parameter values used in metaheuristics,
a sensitivity analysis is conducted to examine the effect
After: 2 2 1 3 1 3 3 2 1 2 of different parameter values on the proposed HBA. Two
parameters, that is, population size 𝑝𝑠 and parameter 𝛼, are
Figure 4: An individual with ten aircrafts before and after LSI local used to investigate the performance with respect to different
search. values, which are based on the average results obtained from
instances involving from 10 to 50 aircrafts, and each instance
run ten times. The principal figures of merit for comparison
randomly. This process is stopped until the difference of of different parameter values are the average percentage
aircrafts amount on each runway is not more than one. gap 𝐺avg (%) associated with the best solution found and
The loudness 𝐿 𝑖 is updated by (12) the average computational time CTavg in seconds for exact
solution is reached. The percentage gap 𝐺(%) is measurement
criteria referenced [2]. Figures 5 and 6 show the statistical
𝐿𝑡+1
𝑖 = 𝛼 × 𝐿𝑡𝑖 , (12) result, where 𝐺avg is shown on the left 𝑦-axis and CTavg is
shown on the right 𝑦-axis and different parameter values (𝑝𝑠
and 𝛼) are listed in on the horizontal 𝑥-axis.
where 𝛼 is a constant and initial value of 𝐿1𝑖 ∈ (1, 2). If a
The performance of the HBA in terms of different
random number is less than its loudness 𝐿𝑡𝑖 , then the RB
population numbers is shown in Figure 5. It shows that better
operation is performed; otherwise, each aircraft is assigned
results can be obtained with a larger population. However, it
to a random runway.
does not improve significantly when the value of 𝑝𝑠 is equal
The iterative process is repeated until the termination
to or greater than 10. On the other hand, the computational
criterion is met; Algorithm 1 shows the pseudo-code of HBA
time increases steadily when the population rises.
based on the framework of bat algorithm for ALP. The lines
Figure 6 illustrates the performance effect of the HBA
1–3, the bat population and other parameters are initialized
with a decreasing 𝛼 value. It indicates that the average
(the initialization of population uses a method described
solution quality deteriorates when the value of 𝛼 decreases
in Section 3.3), and these initial individuals are evaluated
gradually. In some cases, increasing the value above 0.9 may
(Section 3.2 detailedly described this procedure). In lines 5–
actually worsen the average objective value. This is probably
9, the individuals are updated by selecting frequency (this
because the search is extremely random and RB operation is
part corresponds to the position updating of BA; this method
carried out rarely.
is described in Section 3.4), and these new individuals
are evaluated. Lines 10–15 show the local search and these
solutions generated by local search are evaluated. In lines 16–
4.2. Comparisons of Results. Based on the results of the
24, the runway balance operation is carried out. Lines 25-26
sensitivity analysis, the parameter values are set in the
are the judgment of termination criterion and the output of
proposed HBA (i.e., 𝑝𝑠= 10, 𝛼= 0.9) for comparison with
results.
other algorithms for solving the ALP with multiple runways
that were proposed and tested to be valid in previous studies.
4. Simulation Results and Comparisons The computational results obtained are shown in Tables 1 and
2. In Tables 1 and 2, for each problem, the instance (Ins);
The simulation experiment is extensively investigated by a lar- the number of aircrafts (𝑛); the number of runways (𝑚);
ge number of benchmark instances; these well-studied pro- the value of optimal solution (𝑉opt ); the value of the best-
blems are taken from the web OR-Library (last update: June known solution (𝑉best ) if the optimal solution is not known;
2012, http://people.brunel.ac.uk/∼mastjjb/jeb/info.html), a the objective function value (𝑂) obtained and the percentage
reference site which contains detailed information regarding gap (𝐺%) associated with the best solution found over the 15
a large number of benchmark instances. In this paper, the replications and the average execution time (𝑇) in seconds for
whole 13 instances from OR-Library are selected; these 15 replications. Note that in order to compare the results, the
instances have been widely used as benchmarks to certify the 𝑠𝑖𝑗 is set zero in accordance with the previous literature, where
performance of algorithms by many researchers [2, 9]. 𝐺 is calculated on the basis of 𝑉opt or 𝑉best ; 𝐺 = 100 ∗ (𝑂 −
All computational experiments are conducted with MAT- 𝑉opt )/𝑉opt . One complication is that for some problem where
LAB 2012a on a 3.0 GHz Athlon PC with 2.0 GB memory. 𝑉opt or 𝑉best is zero, the percentage gap is defined as zero if and
There are two kinds of termination criterion for different only if the best solution found as zero is also zero, otherwise
instances. For instances where the value of optimal (exact) it is undefined (nd).
solution (𝑉opt ) is known, the algorithm is repeated until Table 1 presents results for a set of instances involving
the objective function value is equal to the 𝑉opt ; if the from 10 to 50 aircrafts. SS is scatter search, the bionomic
objective function value is greater than the 𝑉opt when the algorithm is marked as BA1 [2], and IACA is improved
maximum generation 𝑡max (= 200) is met, then the algorithm ant colony algorithm [8]; heuristic is an effective heuristic
also is terminated. For instances where the value of 𝑉opt is algorithm in [1], FCFS is first-come first-served, and the result
not known, the termination criterion is set as maximum is reference [2]. According to Table 1, the percentage gap
generation 𝑡max = 1000. 𝐺 of SS and BA1 is better than HBA; however, the average
6 Journal of Applied Mathematics

(1) Initialize the ps, 𝑡 = 1, bat population and other parameters;

(2) Construct initial bat population; // (3.3 Initialization construction)
(3) Assign landing time and evaluate each individual; // (3.2 Landing time assignment)
(4) repeat
(5) for 𝑖 = 1:ps do
(6) Determine frequency 𝑓;
(7) Update each bat individual;
(8) end
(9) Assign landing time and evaluate each individual; // (3.2 Landing time assignment)
(10) if rand > 𝑅(𝑡) then
(11) Carry out swap local search operation;
(12) else
(13) Carry out LSI local search operation;
(14) end
(15) Assign landing time and evaluate each individual; // (3.2 Landing time assignment)
(16) Compute loudness of each individual by (12);
(17) for 𝑖 = 1:ps do
(18) if rand < 𝐿 𝑖 then
(19) Carry out RB operation; // Runway balance operation
(20) else
(21) Assign each aircraft to a random runway;
(22) end
(23) end
(24) Assign landing time and evaluate each individual; // (3.2 Landing time assignment)
(25) until 𝑡 = 𝑡max
(26) Output result and plot

Algorithm 1: The pseudo-code of HBA for ALP.

5.25 1.2 5.4 0.6

Average computational time (s)

Average computational time (s)

5.2 1
Average percentage gap (%)

Average percentage gap (%)

5.15 0.8
5.2 0.5
5.1 0.6

5.05 0.4

5 0.2 5 0.4
0 5 10 15 20 25 30 35 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3
Population size Parameter 𝛼

Gavg Gavg
CTavg CTavg
Figure 5: The effect of the value of population size 𝑝𝑠 with respect Figure 6: The effect of the value of 𝛼 with respect to relative
to relative computational time and average percentage gap. computational time and average percentage gap.

execution time 𝑇 of SS and BA1 is longer than the 𝑇 of heuristic. Comparing with IACA and FCFS, it is evidently
HBA and the average execution time of HBA only expends shown that the HBA is effective.
0.43 seconds; by contrast, the SS and BA1 expend 6.4 and Table 2 presents results for a set of instances involving
7.7 seconds, respectively. The comparison of results between from 100 to 500 aircrafts. The termination criterion is set as
HBA and heuristic shows the solutions obtained by HBA are maximum generation 𝑡max = 1000. The IACA and heuristic
superior to the solutions obtained by heuristic. Meanwhile, did not select these instances for testing. From Table 2, even
the average execution time is approximate between HBA and for the larger instances, the HBA found that optimal solutions
Journal of Applied Mathematics 7

Table 1: The comparisons of computational results for first group instances.

HBA SS BA1 IACA Heuristic FCFS

Ins 𝑛 𝑚 𝑉opt
𝑂 𝐺 𝑇 𝐺 𝑇 𝐺 𝑇 𝐺 𝑇 𝐺 𝑇 𝐺
2 90 90 0 0.08 0 2.4 0 4.5 0 0.1 0 0.1 0
Airland1 10
3 0 0 0 0.11 0 3.9 0 3.4 0 0.0 0 0.1 0
2 210 210 0 0.09 0 4.5 0 4.9 0 0.6 0 0.1 9.52
Airland2 15
3 0 0 0 0.10 0 4.6 0 4.3 0 0.4 0 0.1 0
2 60 60 0 0.09 0 4.8 0 5.8 0 11.3 0 0.1 116.67
Airland3 20
3 0 0 0 0.10 0 6.2 0 6.3 0 9.3 0 0.2 nd
2 640 640 0 0.55 0 5.2 0 5.5 0 7.9 0 0.1 0
Airland4 20 3 130 130 0 0.14 0 4.6 0 5.7 0 6.4 0 0.1 0
4 0 0 0 0.14 0 5.6 0 5.2 0 5.8 0 0.2 0
2 650 890 36.92 1.44 0 5.0 3.08 6.1 12.31 2.4 64.62 0.1 16.92
Airland5 20 3 170 170 0 0.16 0 5.4 0 4.3 0 7.0 41.18 0.1 5.88
4 0 0 0 0.21 0 5.6 0 6.8 0 3.2 0 0.2 nd
2 554 636 14.80 1.61 0 7.0 3.61 10.1 51.08 3.8 59.21 0.1 59.21
Airland6 30
3 0 0 0 0.30 0 5.4 0 8.7 0 12.2 0 0.2 0
Airland7 44 2 0 0 0 0.09 0 11.8 0 12.4 0 55.2 0 0.2 0
2 135 180 33.33 2.01 0 12.1 0 19.6 22.22 168.1 88.89 0.2 425.93
Airland8 50
3 0 0 0 0.16 0 13.9 0 18.1 nd 108.2 0 0.6 nd
Average 5.00 0.43 0 6.4 0.39 7.7 5.35 23.6 14.93 0.2 45.30

Table 2: The comparisons of computational results for second group instances.

HBA SS BA1 FCFS

Ins 𝑛 𝑚 𝑉best
𝑂 𝐺 𝑇 𝐺 𝑇 𝐺 𝑇 𝐺
2 452.92 499.49 10.28 16.0 5.67 24.3 54.73 48.7 172.60
Airland9 100 3 75.75 77.03 1.69 16.5 0 39.0 87.46 46.6 342.81
4 0 0 0 16.6 0 33.6 nd 43.9 nd
2 1288.73 1407.45 9.21 18.6 7.87 60.8 25.95 84.5 103.47
3 220.79 224.13 1.51 21.0 8.88 66.8 195.88 80.3 552.16
Airland10 150
4 34.22 34.22 0 20.6 16.74 64.7 292.40 78.8 3473.49
5 0 0 0 23.1 0 60.7 nd 76.2 nd
2 1540.84 1673.95 8.64 23.2 9.19 95.9 38.54 128.7 129.80
3 280.82 280.64 −0.06 26.1 21.59 102.1 290.09 120.3 764.25
Airland11 200
4 54.53 54.53 0 27.4 2.77 99.3 474.47 116.8 3947.88
5 0 0 0 27.2 0 95.6 nd 115.8 nd
2 1961.39 2482.26 26.56 28.3 18.80 126.6 50.18 183.5 137.42
3 290.04 243.79 −15.95 31.5 17.48 145.4 198.01 171.0 903.97
Airland12 250
4 3.49 2.44 −30.09 33.3 271.63 144.5 13216.91 168.8 70752.44
5 0 0 0 34.6 0 138.6 nd 166.2 nd
2 5501.96 5184.06 −5.78 58.0 3.72 383.6 37.47 537.9 56.91
3 1108.51 755.15 −31.88 60.7 1.98 456.0 182.69 515.8 462.60
Airland13 500
4 188.46 90.03 −52.23 63.7 22.98 441.3 1186.81 497.7 2027.94
5 7.35 0 −100 65.9 0 442.1 22308.44 488.7 52628.71
Average −9.37 34.0 21.54 159.0 2576.00 193.2 9097.10

in several cases are prominent. We can clearly find that the and is much greater than the 𝐺 of HBA. On the other hand,
percentage deviation 𝐺 from the best-known solutions is the average execution time 𝑇 expended by HBA is much lesser
negative, which indicates the solutions found by HBA are than the average execution time 𝑇 of SS and BA1 expended.
better than the best-known solutions. The average percentage Figure 7 illustrates the performance of the HBA, the
gap 𝐺 of HBA is −9.37% for this group of instances; however, SS, the BA1, the IACA, and the heuristic with respect to
the average percentage gap 𝐺 of SS, BA1, and FCFS is positive the computational time. HBA and heuristic perform almost
8 Journal of Applied Mathematics

600 [2] H. Pinol and J. E. Beasley, “Scatter search and bionomic

Average execution time (s)

500 algorithms for the aircraft landing problem,” European Journal

of Operational Research, vol. 171, no. 2, pp. 439–462, 2006.
400
[3] J. A. Bennell, M. Mesgarpour, and C. N. Potts, “Airport runway
300 scheduling,” Annals of Operations Research, vol. 204, pp. 249–
200 270, 2013.
[4] A. R. Brentnall and R. C. H. Cheng, “Some effects of air-
100
craft arrival sequence algorithms,” Journal of the Operational
0 Research Society, vol. 60, no. 7, pp. 962–972, 2009.
0 5 10 15 20 25 30 35
[5] M. Wen, Algorithms of Scheduling Aircraft Landing Problem,
Instance no. Technical University of Denmark, DTU, Lyngby, Denmark,
HBA IACA 2005.
SS Heuristic [6] N. Bäuerle, O. Engelhardt-Funke, and M. Kolonko, “On the
BA1 waiting time of arriving aircrafts and the capacity of airports
with one or two runways,” European Journal of Operational
Figure 7: The computational time of each test case. Research, vol. 177, no. 2, pp. 1180–1196, 2006.
[7] X.-B. Hu and E. Di Paolo, “An efficient genetic algorithm
the same on instances up to 50 aircrafts. For the larger with uniform crossover for air traffic control,” Computers and
instances up to 500 aircrafts, the computational time of HBA Operations Research, vol. 36, no. 1, pp. 245–259, 2009.
is shortened observably, which also demonstrates that the [8] G. Bencheikh, J. Boukachour, and A. E. H. Alaoui, “Improved
HBA has a faster convergence rate. ant colony algorithm to solve the aircraft landing problem,”
International Journal of Computer Theory and Engineering, vol.
3, no. 2, pp. 224–233, 2011.
5. Conclusions [9] A. Salehipour, M. Modarres, and N. L. Moslemi, “An efficient
hybrid meta-heuristic for aircraft landing problem,” Computers
In this paper, we considered the multiple runways aircraft & Operations Research, vol. 40, no. 1, pp. 207–213, 2013.
landing problem with the objective of minimizing the total [10] S.-P. Yu, X.-B. Cao, and J. Zhang, “A real-time schedule method
deviation of landing time from the target time. In order for aircraft landing scheduling problem based on cellular
to solve the larger instances involving up to 500 aircraft automation,” Applied Soft Computing Journal, vol. 11, no. 4, pp.
and multiple runways, a hybrid metaheuristic based on bat 3485–3493, 2011.
algorithm (HBA, for short) has been implemented. The HBA [11] G. Hancerliogullari, G. Rabadi, A. H. Al-Salem, and M. Khar-
includes a problem-dependent initialization construction, beche, “Greedy algorithms and metaheuristics for a multiple
and several local search operations are integrated into the runway combined arrival-departure aircraft sequencing prob-
framework of bat algorithm. The computational results of the lem,” Journal of Air Transport Management, vol. 32, pp. 39–48,
HBA show that the proposed algorithm is very effective and 2013.
competitive and can obtain solutions with high quality for [12] D. Briskorn and R. Stolletz, “Aircraft landing problems with
instances up to 500 aircrafts in a short time. Moreover, the aircraft classes,” Journal of Scheduling, pp. 1–15, 2013.
landing time assignment of each aircraft is a key for solving [13] X. S. Yang, “A new metaheuristic bat-inspired algorithm,” in
ALP; several excellent assignment strategies need to be Nature Inspired Cooperative Strategies for Optimization (NICSO
presented in our further work; meanwhile, the aircraft take- ’10), pp. 65–74, Springer, Berlin, Germany, 2010.
off problem (ATP) in airport runway scheduling problem also [14] A. H. Gandomi, X. S. Yang, A. H. Alavi, and S. Talatahari,
is our future work. “Bat algorithm for constrained optimization tasks,” Neural
Computing and Applications, vol. 22, no. 6, pp. 1239–1255, 2013.
[15] X. S. Yang and A. H. Gandomi, “Bat algorithm: a novel
Acknowledgment approach for global engineering optimization,” Engineering
Computations, vol. 29, no. 5, pp. 464–483, 2012.
This work is supported by National Science Foundation of
China under Grant no. 61165015, Key Project of Guangxi [16] S. Mishra, K. Shaw, and D. Mishra, “A new meta-heuristic bat
inspired classification approach for microarray data,” Procedia
Science Foundation under Grant no. 2012GXNSFDA053028,
Technology, vol. 4, pp. 802–806, 2012.
Key Project of Guangxi High School Science Foundation
under Grant no. 20121ZD008, the Funded by Open Research [17] J. Xie, Y. Zhou, and H. Chen, “A novel bat algorithm based on
differential operator and Lévy-flights trajectory,” Computational
Fund Program of Key Lab of Intelligent Perception and
Intelligence and Neuroscience, vol. 2013, Article ID 453812, 13
Image Understanding of Ministry of Education of China pages, 2013.
under Grant no. IPIU01201100, and the Innovation Project of
[18] G. Wang, L. Guo, H. Duan, L. Liu, and H. Wang, “A bat algo-
Guangxi Graduate Education under Grant no. YCSZ2012063. rithm with mutation for UCAV path planning,” The Scientific
World Journal, vol. 2012, Article ID 418946, 15 pages, 2012.
References
[1] J. E. Beasley, M. Krishnamoorthy, Y. M. Sharaiha, and D.
Abramson, “Scheduling aircraft landings—the static case,”
Transportation Science, vol. 34, no. 2, pp. 180–197, 2000.
Advances in Advances in Journal of Journal of
Operations Research
Hindawi Publishing Corporation
Decision Sciences
Hindawi Publishing Corporation
Applied Mathematics
Hindawi Publishing Corporation
Algebra
Hindawi Publishing Corporation
Probability and Statistics
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

The Scientific International Journal of

World Journal
Hindawi Publishing Corporation
Differential Equations
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Submit your manuscripts at

http://www.hindawi.com

International Journal of Advances in

Combinatorics
Hindawi Publishing Corporation
Mathematical Physics
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Journal of Journal of Mathematical Problems Abstract and Discrete Dynamics in

Complex Analysis
Hindawi Publishing Corporation
Mathematics
Hindawi Publishing Corporation
in Engineering
Hindawi Publishing Corporation
Applied Analysis
Hindawi Publishing Corporation
Nature and Society
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

International
Journal of Journal of
Mathematics and
Mathematical
Discrete Mathematics
Sciences

Journal of International Journal of Journal of

Hindawi Publishing Corporation Hindawi Publishing Corporation Volume 2014

Function Spaces
Hindawi Publishing Corporation
Stochastic Analysis
Hindawi Publishing Corporation
Optimization
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014