IEEE Computational Intelligence Magazine - May 2022 - 85

citizen perspective focuses on transportation
services and planning for citizens.
Finally, the last column of Fig. 1 shows
the specific intelligent transportation
application scenes in smart cities.
The contributions of this paper are as
follows: 1) A two-layer taxonomy is proposed
to classify related studies on EC
for intelligent transportation in smart
cities; 2) Related studies are reviewed in
detail based on the two-layer taxonomy;
3) Certain future research directions and
open issues are discussed. This paper
hopes to provide readers with a comprehensive
overview of EC for intelligent
transportation in smart cities and promote
more follow-up research.
The remainder of the paper is organized
as follows. Section II presents
background knowledge about intelligent
transportation in smart cities and
EC algorithms. Sections III-V review
related studies on land, air, and sea
transportation, respectively. Section VI
discusses future research directions
and open issues. Finally, Section VII
draws conclusions.
II. Background
Transportation is a foundational component
in modern cities. The supply of daily
necessities, citizen travel, and commercial
activities all require the support of transportation.
In recent years, the technology
of the Internet of Things (IoT) has
developed rapidly, which enables the collection
of transportation data from multiple
sources (e.g., citizens, vehicles, and
infrastructure). By analyzing these data,
efficient planning and management
schemes can be generated, making
transportation more intelligent in smart
cities. There are four essential objectives
of intelligent transportation: maximizing
transportation efficiency (e.g.,
speeding up transportation, reducing
traffic congestion), maximizing economic
profit, minimizing environmental
pollution, and maximizing benefits
to citizens. Optimization problems of
intelligent transportation in smart cities
often involve large-scale data and complex
features. EC algorithms that have
powerful search efficiency have been
widely used to solve these problems.
The application of EC algorithms for intelligent land
transportation is classified into the following seven
categories: vehicular network, traffic light, electric vehicle
infrastructure, metro transportation, route scheduling, car
service, and trip recommendation.
Therefore, this paper focuses on related
studies on EC for intelligent transportation
in smart cities.
EC is a group of optimization algorithms
inspired by natural phenomena or
social behaviors. Genetic algorithm (GA),
particle swarm optimization (PSO), ant
colony optimization (ACO), and differential
evolution (DE) are typical and
widely used EC algorithms. GA [10] and
DE [11] are inspired by the crossover and
mutation of chromosomes. PSO [12] and
ACO [13] are inspired by the foraging
behavior of birds and ants, respectively.
Based on their search behavior, GA and
ACO are more suitable and easier to
apply in solving discrete optimization
problems, while PSO and DE may be
better choices for continuous optimization
problems. Generally, an EC algorithm
first generates an initial population,
in which each individual represents a
candidate solution of the optimization
problem. Then, a problem-dependent
process to evaluate each individual's fitness
(i.e., the quality of its corresponding
candidate solution) is performed. Next, a
reproduction process that generates or
updates individuals biased by fitness based
on the Darwinian notion is performed,
which is the key component of an EC
algorithm. The algorithm executes the
fitness evaluation and reproduction process
iteratively until the termination condition
(e.g., the maximum number of
function evaluations) is met. Finally, the
algorithm outputs the optimization result.
III. EC for Intelligent
Land Transportation
Land transportation is the most basic
and extensive mode of transportation,
particularly in citizens' daily travel and
short-distance goods transportation.
Land vehicles include cars, trucks/lorries,
and trains. The application of EC
algorithms for intelligent land transportation
is classified into the following
seven categories: vehicular network, traffic
light, electric vehicle infrastructure,
metro transportation, route scheduling,
car service, and trip recommendation.
A. Intelligent Vehicular Network
The vehicular network [14] is a fundamental
component of an intelligent
transportation system in smart cities.
With the development of IoT, cloud
computing, edge computing, and 5G
technology, the application of vehicular
networks has become practical. The
communication unit deployed in vehicles
enables vehicle-to-infrastructure
(V2I) and vehicle-to-vehicle (V2V)
communications. An example of a
vehicular network with V2I and V2V
communications is shown in Fig. 2.
Transportation data can thus be transmitted
and gathered for further analysis
to generate planning and management
schemes. The deployment of vehicular
networks includes a large amount of
infrastructure, which requires a large
amount of investment but yields little
profit. Such a project is typically led by a
government; thus, the application of
vehicular networks is considered from
the government perspective. There are
three critical optimization problems in
vehicular networks: infrastructure
deployment, network routing, and
resource scheduling.
First, although vehicles can communicate
with each other, the function of
the vehicular network also relies on the
deployment of certain infrastructure. Lin
and Deng [15] considered the deployment
optimization problem of wireless
sensors and roadside units
(RSUs).
Wireless sensors are used to collect traffic
data and detect environmental events,
while RSUs are a kind of communication
MAY 2022 | IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE 85
IEEE Computational Intelligence Magazine - May 2022

Table of Contents for the Digital Edition of IEEE Computational Intelligence Magazine - May 2022
