Journal of Physics: Conference Series

PAPER • OPEN ACCESS

Research on Flight Trajectory Optimization Based on Quantum Genetic

Algorithm
To cite this article: Yibo Li et al 2020 J. Phys.: Conf. Ser. 1549 022074

View the article online for updates and enhancements.

This content was downloaded from IP address 178.171.52.208 on 30/06/2020 at 13:29

ESAET 2020 IOP Publishing
Journal of Physics: Conference Series 1549 (2020) 022074 doi:10.1088/1742-6596/1549/2/022074

Research on Flight Trajectory Optimization Based on

Quantum Genetic Algorithm

Yibo Lia, Shuaitao Qin* and Lei Jingb

Shenyang Aerospace University, Shenyang, China

*Corresponding author e-mail: [email protected], [email protected],

b
[email protected]

Abstract. Taking the passenger aircraft A320 as the research object, the trajectory
optimization of the aircraft climbing section under wind disturbance is carried out.
Through the analysis of the aircraft's descent process, the kinematics model, fuel
consumption model and atmospheric model of the aircraft are established, and the
objective function and constraints of the problem are obtained. The quantum genetic
algorithm is used to transform the optimal control problem into a nonlinear
programming problem. The optimal trajectory of the aircraft's shortest descent time is
obtained by simulation. The effects of different cost indices on the fall time and fuel
consumption are discussed.

1. Introduction
As an indispensable part in the field of aircraft aviation, trajectory optimization has been widely used
in aircraft flight management and traffic management, and has received more and more attention.
Trajectory optimization is beneficial to improving the flight efficiency of aircraft and minimizing the
cost while meeting the flight requirements; by using optimization techniques, the traffic congestion
problem can be alleviated and the fuel-saving performance of aircraft can be significantly improved So
that people's travel requirements can be met.
A variety of new methods have been developed based on the early Lev Pontryagin principles, and
both direct and indirect methods of aircraft trajectory optimization have been developed. Among them,
the Lev Pontryagin principle and variational method are the theoretical basis of the indirect method.
Under the influence of the aircraft state variable and the motion equation, the extremum of the
performance index function is obtained. The necessary condition of its intermediate optimality can be
satisfied by its optimal solution, but at the same time it also has some disadvantages, that is, it has a
very complicated derivation process of the optimal solution In the process of solving the two-point
boundary value problem, there will be a small convergence region. In recent years, with the rapid
development of Trajectory optimization method, the problem of optimal control has been well solved,
Wang Yingxun, Chen Zongji, Hu Xiaobing [11, 12] and others used genetic algorithm (GA) to avoid
the constrained flight path planning, and Yeonju [13] used genetic algorithm to establish the no-fly
zone for several uavs The Path Planning and cooperative task assignment are realized, and then,
because of the influence of stochastic search, the search efficiency of genetic algorithm will be
reduced under the influence of population degeneration, and then the convergence speed will be
reduced Duan Jiajia [8] and others studied the optimization of the lunar soft landing trajectory by

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
ESAET 2020 IOP Publishing
Journal of Physics: Conference Series 1549 (2020) 022074 doi:10.1088/1742-6596/1549/2/022074

using the ant colony algorithm, but the search time was too long and it was easy to fall into the local
optimal solution, Xu ying [6] through the use of Particle swarm optimization to complete the study of
the aircraft's climb-cruise and descent of the optimal trajectory, wang Wei [9] used the energy state
method to realize and optimize the genetic algorithm and the fuel-saving trajectory of the aircraft.
However, because of the simplicity of the aircraft model, safety is not fully considered in civil aviation
flight the actual civil aviation can not fully use the climb trajectory of the aircraft.
Quantum Genetic Algorithm (QGA) is a new intelligent Algorithm which combines the traditional
Genetic Algorithm (GA) with quantum theory, and is first proposed by Narayanan Han then extends
QGA by superposing quantum states and qubits. QGA is widely used in many fields such as path
optimization, scheduling optimization and signal processing because of its outstanding advantages
such as good population diversity, strong global search ability and fast convergence speed.

2. Basic QGA
With the full integration of GA and quantum computing, a new field of research called Basic QGA has
emerged, which is based on quantum computing properties such as quantum entanglement and
quantum parallelism Compared with GA, QGA has stronger parallel processing ability and faster
convergence speed by using quantum revolving gate and multi-state gene qubit coding to update the
operation.

2.1. Qubit
For a quantum computer, a two-state quantum system acts as the physical medium of the information
storage unit, and thus constitutes the qubit. Unlike a classical bit, a qubit can act simultaneously in a
superposition of two quantum states, such as    0  1 , where（，）is a pair of amplitude
2 2
constants that Satisfy     1 0 and 1 to represent the spin down and up states, respectively
The information in 0 and 1 can be contained in the same quantum bit.
In QGA, a gene can be stored and expressed by quantum bits. The "0" or "1" states can be used in
the gene, or they can be superimposed at will. That is, the gene is no longer able to express certain
information, but contains all of it. Through the operation of the gene, it can express all kinds of
possible information simultaneously. With the use of quantum bit coding, the same chromosome can
be simultaneously expressed in multiple superposition states, which greatly increases the diversity of
2 2
QGA features. We can get better convergence by using quantum bit coding. When  or 
fluctuates at or near 0 or 1, the chromosome can be in a single state of convergence by using quantum
bit coding.

2.2. Quantum Measurements

The goal of quantum measurement can be realized by the measurement of individual population, that
 
is, a group of definite solutions can be calculated p (t )  p1t , p2t  , pnt , in this case, the J solution in
the t population (the measure of the j individual) is represented by p tj , the binary string is its
representation, m is its length, 0 or 1 is its representation at each position Based on the quantum bit
2 2
probability（ it or  it ，i  1,2, m) the measurement process is as follows, the data with interval
0,1 is selected randomly. If its value is above the square of probability amplitude, the measurement
result is 1, otherwise the value is 0.

2.3. Quantum Updates

For evolution operation, Quantum Gate is the executive mechanism of evolution operation, which can
be selected according to specific problems. Nowadays, there are many kinds of quantum gate, and

2
ESAET 2020 IOP Publishing
Journal of Physics: Conference Series 1549 (2020) 022074 doi:10.1088/1742-6596/1549/2/022074

quantum gate should be regarded as the executive mechanism of evolution operation according to the
characteristics of QGA The quantum gate a can be adjusted by the following relation, which can be
used to select the suitable quantum revolving gate U（）. That is:
cos  sin  
U ( )  
 sin  cos 
Where the rotation angle is represented by  . Its symbol and size need to be preset first, the
following relationship for the update process:
 i  i  cos i  sin i   i 
    U (i )      sin  cos i   i 
 i  i  i
T
In the formula,  i i T and  i i  respectively represent the probability amplitude of the
 
ith qubit revolving gate of chromosome before and after updating.
     i cos i  i sin i
From the above formula,  i i are respectively:  i
i   i sin i  i cos i
2 2 2 2 2 2
So:  i   i   i   i  1 can see that after the transformation  i   i is still 1

3. Problem Description
Kinematic model of aircraft
After considering the mass of the aircraft, the three-degree-of-freedom motion mathematical model
d
（R  H）cos  V cos  cos   Wx
dt
d
（R  H）  V cos  sin   Wy
dt
dH dm
 V sin  ,   Ffuel
dt dt
of the change is established, namely:
dV
m  T cos   D  mg sin 
dt
d
mV cos   L sin 
dt
d
mV  T sin   L cos   mg cos 
dt
In the formula, T, L, D are used to express the thrust, lift and drag of the engine respectively, and
the fuel consumption per unit, the component of the wind field along the longitude direction and the
component of the wind field along the latitude direction are represented by F fuel , Wx and Wy .
When the throttle opening is fixed, the engine thrust is also changing during the flight of Mach and
the flight altitude, which can be expressed by the following relationship:
T  T0 ( H , Ma) 
In the upper formula, the throttle opening is indicated by  .

3.1. Objective Function

In the case of Flight Management Systems, the implementation of vertical navigation functions
requires the calculation of the main parameter, the cost index, which refers to the ratio of the cost
incurred under the effect of flight time to the fuel cost, namely:

3
ESAET 2020 IOP Publishing
Journal of Physics: Conference Series 1549 (2020) 022074 doi:10.1088/1742-6596/1549/2/022074

Ci  Ctime / C fuel
In the formula: The cost index and the cost index are represented by Ci and Ctime respectively; the
cost of the fuel consumed by the aircraft 1B is represented by C fuel ; the optimization of the aircraft
trajectory is mainly to reduce the direct cost according to different cost index, the optimization is
realized, and then the different state information of aircraft at different time is obtained.

3.2. Normalization
In the process of solving the optimization problem, if we want to improve the convergence speed
of the algorithm, we need to deal with the aircraft equation by normalization. A dimensionless
physical quantity is defined as:
V  V / kV , R  R / k R
H  H / k H , m  m / km
Thus, under the constraints of aircraft motion model, boundary conditions, state control variables
and intermediate processes, the optimization problem of aircraft trajectory can be transformed into an
optimal control problem the objective function of the latter is mainly the arrival time and the minimum
time cost, namely:
tf tf
min J ( X (t ),U (t ), t )   Ffuel (t )dt  Ci   dt
t0 t0

X  f ( X (t ),U (t ), t )
X min  X  X max
U min  U  U max
 ( X (t0 ), t0 , X (t f ), t f )  0
C ( X (t ),U (t ), t )  0

4. Trajectory optimization of climbing section

According to the quantum genetic algorithm and the actual trajectory optimization model as shown in
the climb trajectory optimization algorithm flow:
(1) Population Q (t0 ) is initialized by the coevolutionary strategy of Niche Coevolution, n
chromosomes are generated randomly, and all chromosomes are encoded by Qubit.
(2) Measure each individual in the initial population Q (t0 ) one by one, so as to obtain the
corresponding definite solution P t0  .
(3) Evaluate the various deterministic solutions.
(4) Keep track of individual fitness and optimization.
(5) To judge whether it can be finished, if it can satisfy the condition, it can be quit, otherwise it
needs to be further calculated.
(6) All individuals in population Q t  are measured one by one and the corresponding definite
solution is obtained.
(7) Evaluate the various deterministic solutions.
(8) With the help of the dynamic adjustment mechanism of the rotation angle of the quantum
revolving gate U t  , the selection and crossover are used to realize the relative operation to the
individual, thus the acquisition of the new population Qt  1 can be realized.
(9) Keep track of the optimal individuals and their fitness.
(10) Add the number of iterations to 1 and return to step 5.

4
ESAET 2020 IOP Publishing
Journal of Physics: Conference Series 1549 (2020) 022074 doi:10.1088/1742-6596/1549/2/022074

4.1. Simulation and Analysis

Taking the Airbus A-320 as an example, this paper completes the ambush optimization of the climbing
section by using the algorithm mentioned above. The initial speed of the plane is 10m/S, the plane
climbs gradually from the bottom to the altitude of 10000m, the terminal speed is 280m/S, and the
constraint condition is 0 ≤ 25 0≤ Ma ≤0.85,0≤ n ≤1.75, population size and evolutionary Algebra are n
300 and m 150 respectively, crossover probability 0.8, mutation probability 0.05.
Simulation Results:

Figure 1. Climb flight path.

Figure 2. Mach Number, track angle, flight altitude with time curve.

Figure 3. Curves of angle of attack and overload with time.

5
ESAET 2020 IOP Publishing
Journal of Physics: Conference Series 1549 (2020) 022074 doi:10.1088/1742-6596/1549/2/022074

Figure 4. Curves of thrust and mass versus time.

Figure 5. Evolution and fitness curve.

4.2. Result Analysis

As you can see from the oil diagram, after the trajectory optimization of the climb phase of the civil
aircraft is completed by quantum genetic algorithm, the requirement of two climbs can be met Fig. 2
illustrates in detail the variation of Mach number, altitude and track angle of the passenger aircraft
during the crawling process. The speed of the aircraft can be significantly increased in the range of
110-150s, this area happens to be in the process of the first level flight of the aircraft, thus
accumulating climbing power to prepare for the second climb. According to the engine curve in figure
4, the overall characteristics of the engine decrease with the increase of the aircraft altitude, so the first
level flight can improve the efficiency of the engine. As can be seen from Fig. 3, the overload and
angle of attack control variables of the passenger will change in the range of [0, 1.5 G] when the
aircraft is climbing, so as to satisfy the comfort requirements of the passenger Aircraft generally need
to pass through 270 seconds of crawling time, need to consume 200kg of fuel, to meet the relevant fuel
consumption standards.
Fig. 5 shows the optimization process based on quantum genetic algorithm, that is, it reflects the
change law of the population's optimal individual fitness, and finally the optimal individual fitness
tends to be stable and has good convergence, the results show that the algorithm is very effective.

5. Conclusion
In this paper, the application of quantum genetic algorithm (QGA) is used to complete the research on
the climb flight path optimization of civil aviation passenger aircraft. Taking Airbus A-320 as an
example, the effectiveness of QGA is verified Quantum genetic algorithm (QGA) can satisfy all the
technical constraints, and also can achieve the climb requirements of civil aircraft, which shows that
QGA can effectively solve the climb trajectory optimization problem of civil aircraft.

6
ESAET 2020 IOP Publishing
Journal of Physics: Conference Series 1549 (2020) 022074 doi:10.1088/1742-6596/1549/2/022074

References
[1] Lin Jun, he Yingzi, Huang Panxing. Reentry trajectory planning of hypersonic vehicle with
thrust based on improved piecewise Gauss pseudo-spectral method [J / Ol]. Control Theory
and applications: 1-10 [2019-10-10]. HTTP: Kns Ki. netkcmscdetail
44.1240.TP20190318.1057.050. HTML.
[2] Zhang Yaping, Sun Peihua, Li Yuhui, Liu Yanbin. Trajectory optimization of hypersonic
vehicle based on improved Pigeon Swarm Algorithm [J]. Flight Mechanics, 2017, 35 (04):
60-64.
[3] Taukin, Sun Qinglin, Zhu Erlin, Chen Zengqiang, he yingping. Homing trajectory planning of
parafoil system based on quantum genetic Algorithm [J]. Journal of Harbin Engineering
University, 2016, 37 (09): 1261-1268.
[4] Xue Pengfei, Liu Xiaoxiong, Li Chuang, Ma Qingyuan. Optimization of aircraft descent
trajectory based on Gauss pseudo-spectral method [J]. Flight Mechanics, 2016, 34 (03): 34-
38.
[5] Fish Jiaxin, Li Gang, Li Dongtao, Li Wenjun. Application of improved quantum genetic
algorithm in UAV path planning [J]. Computer Simulation, 2015, 32 (05): 106-109 + 131.
[6] Xu Ying, Liu Xing, Sun Xiaoyang. Particle swarm optimization based cost-of-flight trajectory
optimization [J]. Aircraft design, 2011, 31 (01): 32-35.
[7] Wu Di, Wang Feng, Yang Shengyao. Research on flight trajectory optimization based on chaos
immune genetic Algorithm [J]. Journal of Wuhan University of Technology, 2010, 32 (20):
147-150 + 155.
[8] Duan Jiajia, Xu Shijie, Zhu Jianfeng. Optimization of lunar soft landing trajectory based on ant
colony Algorithm [J]. Journal of Astronautics, 2008 (02): 476-481 + 488.
[9] Wang Wei, Ning Dong Fang, Zhang Jin. Optimization of aircraft fuel-saving trajectory based on
energy state method and its genetic Algorithm Implementation [J]. Measurement and control
technology, 2006 (01): 56-58 + 65.
[10] Wang Hongzhi, Wang Zhengping, Zhu Tiefu. Research on solving aircraft optimal climb
problem based on genetic Algorithm [J]. Journal of Missiles and guidance, 2005 (S5): 624-
625.
[11] Hu Xiaobing, Wu Shufan and Jiang Ju. Ga based dynamic optimization algorithm for horizontal
flight path of Civil Aircraft [J]. Aeronautical Computing Technology, 2001 (02): 36-41.
[12] Young-hoon Wong, joong-ki Chan. Constrained flight path planning based on genetic algorithm
(Ga) [J]. Journal of Beijing University of Aeronautics and Astronautics, 1999 (03): 113-116.
[13] Yeonju Eun and Hyochoong Bang. Cooperative Task Assignment and Path Planning of
Multiple UAVs Using Genetic Algorithm [J]. AIAA Infotech@Aerospace 2007Conference
and Exhibit 7-10, May 2007, Rohnert Park, California. AIAA 2007-2982.