Help the Genetic Algorithm to Minimize the Urban Traffic on Intersections

International Journal of Research in Computer Science
eISSN 2249-8265 Volume 4 Issue 4 (2014) pp. 9-19
www.ijorcs.org, A Unit of White Globe Publications
HELP THE GENETIC ALGORITHM TO MINIMIZE THE
URBAN TRAFFIC ON INTERSECTIONS
Dadmehr Rahbari
University of Applied Science and Technology, IRAN
Email: d_rahbari@yahoo.com
Abstract: Control of traffic lights at the intersections
of the main issues is the optimal traffic. Intersections
to regulate traffic flow of vehicles and eliminate
conflicting traffic flows are used. Modeling and
simulation of traffic are widely used in industry. In
fact, the modeling and simulation of an industrial
system is studied before creating economically and
when it is affordable. The aim of this article is a smart
way to control traffic. The first stage of the project
with the objective of collecting statistical data (cycle
time of each of the intersection of the lights of vehicles
is waiting for a red light) steps where the data
collection found optimal amounts next it is.
Introduced by genetic algorithm optimization of
parameters is performed. GA begin with coding step
as a binary variable (the range specified by the initial
data set is obtained) will start with an initial
population and then a new generation of genetic
operators mutation and crossover and will Finally,
the members of the optimal fitness values are selected
as the solution set. The optimal output of Petri nets
CPN TOOLS modeling and software have been
implemented. The results indicate that the
performance improvement project in intersections
traffic control systems. It is known that other data
collected and enforced intersections of evolutionary
methods such as genetic algorithms to reduce the
waiting time for traffic lights behind the red lights and
to determine the appropriate cycle.
Keywords: Urban Traffic, Petri Net, Genetic
Algorithm.
I. INTRODUCTION
The Goal of this project is to obtain the optimal
cycle traffic lights at the intersection. To find time for
each cycle of the traffic lights are red, yellow and
green is the average number of waiting cars at behind
of a red light. Traffic Modeling with Petri nets in CPN
TOOLS software simulations has been conducted.
Data collection was performed at various times as the
objectively reasonable sample has been collected. The
results of the genetic algorithm is used to optimize the
design so that the chromosome based on 14
parameters (5 for the intersection lighting cycle time,
cycle time 5 to the second intersection lights, light, on
the intersection of the first two, to wait for an average
of vehicles and two average cars waiting at a red light
to the second intersection), coding (how to convert
binary digits the actual amount of decimal and vice
versa), mutation operators (to create diversity in the
population), crossover (to converge to the optimal)
fitness function (assessment of chromosomes in
population) and elitism selection (to select the number
of inhabitants to move to the next generation) are
described in detail in the relevant sections.
All system controls the timing of traffic lights after
crossing lights as decision variables in the objective
function in order to meet minimum system
requirements. The decision variables include the
proportion of green time of each phase, cycle length,
number of phases, the offset is. Various objective
functions such as energy loss, reducing delays,
reducing the number of vehicles stopped at the
intersection, reduce the length of queues, increase
safety, etc., is considered the beginning of the traffic
light control system for different types of systems
have been created under as controls when the controls
have been categorized prematurely[1]. Other divisions
such as independent monitoring, control and
accountability varies depending on time of day and
traffic controls exist. This is set in Table 1 is
presented. Variable three generations of light-
controlled junctions in a variety of computer programs
with a view to expanding the use of computer
processors are included. This controls how a software
www.ijorcs.org

10 Dadmehr Rahbari
system, based on information received from
computed vehicles and pedestrians, and other inputs
to recognition of designer (eg weather conditions) are
determined. Table 2 presents a set of best known and
most widely used of these systems are discussed [2].
In [3]a traffic control signal to adjust their schedule.
This is a discrete time model optimized for signd areas
marked provides a real-time embedded controller and
the controller signal timing based on technical
constraints, physical and specifies operations. In
[4]proposed a system for wireless traffic light
controller both manual mode and automatic. Manual
control of traffic lights to change with the traffic
police to provide With the push of a button on the
streets for a Green Mark provides. In automatic mode
traffic light controller board, LED sequence according
to a predetermined pattern, and latency changes that
will help the police to be able to dynamically change
the flow of traffic. In [5] a fuzzy logic based smart
traffic light simulator design presented. Traffic
junction simulator hardware is to overcome the
problems of working in a real environment to develop
and test the controller performance. Simulation results
using traffic lights with light controller based on fuzzy
logic is compared for constant and controller. The
results showed that the controller reduced the waiting
time at red lights. In [6] the effects of traffic lights at
intersections to optimize evacuation routes were
investigated. They believed that an intersection with
traffic lights consist of a series of sketches of the
windows when the term is assigned. Considering the
delay and the capacity allocated to each intersection in
the direction of discharge flow path, Compound
labeling algorithm for the minimum cost network
traffic is proposed. In numerical examples with
different latencies compared Evacuation Routes which
showed that the proposed method can be a good result
considering the time of his discharge in path
optimization.
In [7] of the network traffic signal control strategy
is presented. They propose a model to simulate the
UTNS where the road network topology dynamically
directed graph is given by CTM cell transfer model of
traffic flow and transport modeling is and then a state
feedback control law is designed to achieve stable
agreements under which the proposed control system.
In [8] a distributed and adaptive traffic signal
control to provide a realistic traffic simulation. This
method of traffic control algorithms based
observations suggest that the traffic generated by
sensing devices existing local communication between
traffic lights and driving convenience. This method
allows features to be updated frequently to adapt to
the current traffic demand. Given the proposed
adaptive system simulation environment CUMO,
overall network performance garnered a higher level.
Several applications of hybrid Petri Net is presented in
the various articles. The oil industry and oil pipelines,
in [9] to model the transmission lines and oil storage
tanks using HPN described in[10] a short time to
transport crude oil to refineries in Port through
transmission lines is presented. Also in [11] for the
automated online fault monitoring using hybrid Petri
net is done. For intercity traffic problem, in [12]
traffic and car speed in a road tunnel with HPN
models and interventions to prevent accidents in the
tunnel is presented [13]a model based on hybrid Petri
net for the Category movement intercity trains and
avoid collisions with each other and the train
crossings cross- other vehicles on rail tracks designed.
As well as numerous works by using hybrid Petri nets
for modeling, control and optimization of network
traffic have been done. For example, in [14] model of
an urban intersection using hybrid Petri nets have
gained. In this model, the associated status indicator
corresponding to a given area's streets, neighborhoods
have been applied consistently 4 to isolate areas of our
neighbors. In [15] a typical intersection using a
method similar to [16] have modeled. In this paper,
the optimal timing of traffic lights at each phase of the
streets leading to the intersection number of vehicles
expected to be calculated. In the next cycle, the
information flows from the streets into the streets
leading to the intersection of the current cycle, is
obtained.
In [17] the intersection of the model with hybrid
Petri nets attached to each block consists of a hybrid
Petri net is simulated. The block consists of hybrid
Petri nets as a time delay which represents the
movement of vehicles between the two intersections
leading to the street. In [18] In addition to modeling
the hybrid Petri nets, with the introduction of the
concept as a way to obtain the shock wave, the flow
discontinuities between flows vehicles when there are
such intersections is presented.
II. THE PETRI NET
Petri net is a powerful tool for modeling
concurrency and provides more context than queuing
networks. In addition to the structure and behavior of
Petri nets are a formal, graphical display that is why
www.ijorcs.org

Help the Genetic Algorithm to Minimize the Urban Traffic on Intersections 11
they are modeled by easy. One reason for the success
of Petri nets is their simplicity, but the simplicity is
sometimes difficult to model complex systems. Many
features have been added to the basic model of Petri
nets to the modeling, it will increase and can be used
in various fields. Petri nets graphically displayed to
illustrate the theoretical concepts of Petri nets is very
useful. Petri net structure of Petri nets to provide a
graph in which there are two types of nodes. The
barriers in the form of a circle (O) and line (|) indicate
that abrasive tracks are places and marker lines. These
places the tracks are connected to each other by
transition. Transfer from one place to another place,
indicating that it will be passed as output and if it
draws an arc from a place to a transition indicates that
it will be the entrance to the pass[19].
One of the concepts that should be considered in
the evaluation of the performance of talk time.
Colored Petri nets in the concept of Global at the time
of the element name defined. It expresses the values
that are at the disposal of the model. This time can be
an integer or a real number representing the discrete
time which represents the time has come. Besides
being include every nut, some referable, can be
applied to any bead that amount of time. This amount
of time is called time stamp. time stamp represents the
first time that the beads can be used.
In a colored Petri net is based on the time when the
bolts are in place and guard the entrance of a transfer
of the transmission is established in this case, the
desired transformation is activated. But for this
transition can be sure that the fire is in a state of
readiness. When transfer is in standby mode, the
transmission input housings beads are smaller or equal
to the current time. For an event that takes r time units
into a colored Petri net model, The corresponding
transition event, which took thousands of stamp seals
for nut production output r time units larger than the
global time at which the transfer occurred. This means
that the pieces produced by r time units are not
available. Time based on colored Petri nets, while the
current model, no transition was not ready, Time to
pull the system down to at least one transition is in
standby mode. Various tools that can support this kind
of network software Design / CPN named. Using the
graphical user interface of the software, can be
colored Petri net models can be easily traced. You can
also use this animation software (stepwise
implementation) of colored Petri net models, Create a
state space model of colored Petri net, to model the
behavior of a user-defined query and create output
files for the model simulation results are also
provided. CPN models are the benefits that can be
mentioned the following:
Due to the hierarchical model can be drawn from
these techniques can be combined to obtain or reuse of
models can be used. For each transition can be defined
as the time it takes to fire the bolts can be given that
any time a character is also, useful tool for evaluating
performance and other issues related to time. Place
combines a natural way to combine hierarchical
colored Petri nets model is. Timing of the transition in
the Petri net: To create a time delay in the transition
from the Transition Place the following steps should
be taken: Defines the color of the closet:
No = with R | G | Y timed;
Define the variable as
Var a: No;
The Place for No and ◊ study of a defined type.
Place the required amount of sample R @ + 5, from 5
to start the transition and the transition in the
Transition can the formula for the passage of time @
10+ write transition the passage of any 10. CPN Tools
for modeling software proud 4 traffic lights in two
intersections are shown in Figure . To design such a
model, we first create an empty Petri nets. Place the
tool bar and then put it on the network by using the
Style tool to the red, green and yellow in it. Transition
between Places to put some transition state. The next
phase of traffic lights cycles using the Arc tool
connects us. We do this for both lights. Then, using
whatever tools Yellow Arc of Transition transitions of
traffic lights using the tool called X that represents the
intersection Place, the connect them. Is required on
each arc is a variable whose type was defined as the
amount of our No. The optimal values of the genetic
algorithm to optimize the lights that have earned the
tool between any two Place in Transition (@ + x), we
write to you indicate a time delay switch the lights
from one state to another.
III. OPTIMIZATION USING GENETIC
ALGORITHMS
Genetic algorithm to simultaneously consider
multiple points of the search space and thus increases
the chances that converge to a local maximum,
decreases. Search in more conventional ways, the
decision rule governing this case serves to move from
one point to another in a matter of searching this way
can have a maximum bite. Because they may
converge to a local maximum. But the whole
www.ijorcs.org

12 Dadmehr Rahbari
Figure 1: Modeling by Petri Net
population of the genetic algorithm (strings) to
produce and test each point individually by combining
quality (content) of spots, a new population, which
includes the improvement of has the form. Apart from
doing a search, consider the same number of points in
the genetic algorithm, which makes it adaptable to
parallel machines because of the evolution of each
point, is an independent process. The genetic
algorithm only requires information on the quality of
the solutions produced by each set of variables. If
some optimization methods require derivative
information or even need to have a complete
understanding of the structure of matter and variables.
Because genetic Algorithm not require such specific
information on the issue is thus more flexible than
most search methods. The genetic algorithm is a
search method, which is to guide the search for
methods of use random selection will vary. Although
the decision as to define methods of accident and
chance, but the search space is not a random walk.
Genetic algorithms are suitable for the crash exploit a
priori knowledge they use to solve the nearly optimal
to quickly reach it.
The coding problem and a fitness function to
determine the barrier population over generations by
using the operators of selection, mutation and cutting,
elitism is to involve the local optimal solutions[20].
Elitism in the replacement of a case is done.
The coding of chromosomes and the problem:
110 ...
111
101 ...011010 ... 101
111 ...
101
x13 x14x11 x12
x6 x7 x8
x9 x10
x1 x2 x3
x4 x5
The
number of
cars
waiting
for the
red lights
at the
crossroad
s TL3 and
TL4 II
TL1 and
TL2
number of
cars
waiting
for a red
light at the
intersectio
n of
5 stages of
the cycle
time of the
second
crossing
lights
5 stages of
the cycle
time of
the first
crossing
lights
Figure 2: Coding Problem in Genetic Algorithms
Because the chromosomes are composed of a
number of genes in the gene is considered a digit
between is zero and 10 . Xi is the number of
parameters (x1, x2, and X14) and 8 bits for each
www.ijorcs.org

parameter is selected. For example, to convert
decimal number from 1 to 8 bits, the bits x1 and
convert the decimal number from 9 to 16, the value of
x2 as well as X14 ..
• 2 lights with different colors
• green (G), orange (O), or red (R)
• Any integer means the time (in seconds) of each
cycle is a light in the intersection.
• The binary chromosome becomes.
A. Fitness function
Fitness Changes in the genetic algorithm as the
fitness function, the objective function is considered.
In this case the fitness function is based on an average
cycle time of traffic lights. This function has 14
parameters which are specified range of values[21].
Linespace function in Matlab software the specified
range is divided into 100 so that the fitness value is
used calculation. This function has a minimum and
maximum value and the boundary value of the
chromosome is randomly generated value between
these two values to be given. Limit values based on
data obtained from the data collection phase of the
two intersections Flowers and knowledge of
governmental requirements. These values include
cycle time lights at two intersections and the number
of cars waiting for the red light is back. Since genetic
algorithms are used to calculate the maximum
amounts referred to reverse the charges.
Definition of fitness function:
Fitness(gen) = ( (x1(gen) + x2(gen) + x3(gen) +
x4(gen) + x5(gen))*(x11(gen) + x12(gen)) +
(x6(gen) + x7(gen) + x8(gen) + x9(gen) +
x10(gen))*(x13(gen)+x14(gen)) ) / 14 ;
Average time to complete cycle lights at both
intersections with the number of cars waiting behind a
red light, according to the following parameters are
determined.
Parameters x1 x2 x3 x4 x5 for the first 5 stages of the
cycle crossing lights, x6 x7 x8 x9 x10 parameters for
the 5 stages of the second intersection lights cycle
time parameters for the number of cars waiting behind
x11 x12 TL0 and TL1 red lights at the intersection of
First and parameters the number of cars waiting for
the red lights x13 x14 TL2 and TL3 are at the second
intersection. The total cycle time for each intersection
lights on top of the total number of cars waiting for
the red lights would be multiplied The total cycle time
for each intersection lights on the total number of cars
waiting for the red light becomes multiplied by the
average of these values is calculated.
B. Selection Operator
The selection of parents in genetic algorithms to
further its chances of reproduction of the members
that have higher fitness. There are several ways to do
this. A technique that is commonly used selection
method using a wheel. The implementation of this
method is as follows: A - the fitness of all members of
the population and thus the suitability of the call stack.
(B) a randomly generated number n, so that it is a
number between zero and total fitness. (C) The first
member of the population that add elegance to the
fitness of the population of the former is larger than or
equal to n restore[22]. Effects of parental choice back
wheel of a parent is randomly selected. Although the
selection process is random, the chance that each
parent is selected is directly proportional to its fitness.
The balance in the number of generations the
genetic algorithm with the lowest fitness of the
disposal and the release of genetic material in the most
graceful members of help. It may be the worst
member of the population can be chosen by the
algorithm (because anyway there is a random element
in the algorithm.) In the population acts of violence,
however, are negligible and the assumption that
members of a generation, the next generation is much
more likely to desorption. However, after many
generations, the members of the population are
excreted. Parents should be careful in the process of
selecting a range of fitness levels should be positive
integers.
C. Crossover operator
The performance of this operator and mutation
operator causes the fibers produced during
reproduction, the parents are different disciplines. In
nature, this function occurs when two parents
exchange portions of their corresponding fields and
genetic algorithms, operator communication,
exchange of genetic material between the parent
sequences to the child (discipline) create there. There
are several types of exchange operator. But the most
famous exchange operator used in genetic algorithm is
a function of a point[23]. The genetic algorithm is the
operator in the manner described below may apply.
www.ijorcs.org

14 Dadmehr Rahbar
Figure 3: A view of the intersection operator
To be able to use this function need to have two
fields. The selection operator is applied to the current
population of the two strands of the double-take then
do a test to determine the probability that the
exchange operator acts on two fields it or not. This
test is done using a heterogeneous coin, this means
that the probability (P crossover) milk and with
probability of crossover line comes from. For
example, if the applied field with a tap on the coin
exchange operator, we assume that we have thrown a
coin into the milk. Then enter the next phase of the
implementation of the exchange operator, we generate
a random number between one and the length of the
string. After determining the location of the integer
that represents the exchange of strings two strands of
the location of this specifies the number of broken and
distal parts are interchangeable with each other. All
parts are separated from each other are connected to
the new string is obtained.
D. Mutation Operator
The operator is also one of the operators of genetic
algorithms and the ability to use genetic algorithms to
find near-optimal solutions increases. Mutations,
random changes in the value field is a special
situation. By applying this operator characteristic that
is absent in the parent population, is created. Because
mutations alter a gene, i.e. if the value is zero, and
vice versa if it is a zero. So why the change is
characterized by a series of premature convergence
and to not be perfect. Because one of the causes of
premature convergence of the population is members
of the same mutation causes the same probability of
being members of the new population is much
reduced. The implementation of this function is
described below.
Fig 4. Mutation Operator
This operator, unlike the exchange operator to
compare two strings needed to cover a range of needs,
after the exchange operator acts on two fields and two
new strings to the operator of mutations to the double-
stranded is applied to either separately. The method is
applied to the individual elements of a string, mutation
testing is done. If this test is successful, the status is
changed from one to zero or from zero to one, and the
so-called mutations. Test the possibility of using
heterogeneous coin with probability (P crossover)
milk and with probability line comes with a coin toss
will be done and if the milk is collected bit value of
the mutation the[24]. As noted above, the probability
for each state of a field test should be performed. In
other words, for every mutation released once the coin
is heterogeneous with regard to the outcome, finds
little bit mutation or a mutation goes no further.
E. Elitism
With regard to a possible value can be a percentage
of the population without mutation and crossover
operators are transferred to the next generation. This
amount is likely to vary depending on the values that
make up the result set. Genetic algorithm performance
is significantly dependent on the different stages of
the skins. For instance, each of the following: improve
the efficiency of the routing issues.
F. Genetic Algorithm Settings
During several performances, the best parameter
values of the genetic algorithm are obtained as
follows. Total population = 20, number of generations
or iterations = 200, length = 80 chromosomes, the
number of parameters: 10, Pm: probability of
mutation equal to 0.02, Pc: probability of crossover
operator is equal to 0.5, elitism rate: 0.02 times the
risk of elitism.
IV. SIMULATION AND RESULTS
To evaluate the proposed approach to data integrity
requirements that is listed below. Assists the
intersection of knowledge and statistical data are
collected in sampling. At each intersection of the 20
www.ijorcs.org

samples taken at different times. Samples so that the
number of cars waiting behind a red light at the
intersection of the two traffic lights at each
intersection with traffic lights there are a total of 4.
The data collected are given in the following tables.
Table 1: Number of cars behind traffic lights at each intersection
Row
Cars at a red
light TL1
By a red light
TL2
The number
of cars at a
red light TL3
The number of
cars at a red
light TL4
Count/Hour Count/Hour Count/Hour Count/Hour
1 33 19:00 25 19:00 35 19:00 5 19:00
2 25 19:05 30 19:05 25 19:05 10 19:05
3 26 19:10 33 19:10 28 19:10 8 19:10
4 36 19:15 40 19:15 21 19:15 12 19:15
5 22 19:20 34 19:20 27 19:20 11 19:20
6 26 19:25 35 19:25 28 19:25 14 19:25
7 30 19:30 40 19:30 29 19:30 16 19:30
8 28 19:35 38 19:35 26 19:35 18 19:35
9 38 19:40 42 19:40 32 19:40 15 19:40
10 35 19:45 40 19:45 35 19:45 22 19:45
11 40 19:50 39 19:50 29 19:50 23 19:50
12 48 19:55 48 19:55 31 19:55 20 19:55
13 46 20:00 46 20:00 38 20:00 28 20:00
14 49 20:05 49 20:05 41 20:05 31 20:05
15 51 20:10 45 20:10 45 20:10 26 20:10
16 58 20:15 51 20:15 48 20:15 33 20:15
17 55 20:20 55 20:20 51 20:20 36 20:20
18 62 20:25 46 20:25 54 20:25 33 20:25
19 54 20:30 50 20:30 55 20:30 38 20:30
20 60 20:35 54 20:35 58 20:35 37 20:35
Table 2. Line Cycle in intersections
Light Cycle
1
Light Cycle
2
Light
Cycle 3
Light
Cycle 4
TL1 Sec TL2 Sec TL3 Sec TL4 Sec
G 38 G 40 G 60 G 22
Y 5 Y 5 Y 5 Y 5
R 50 R 40 R 30 R 66
www.ijorcs.org

16 Dadmehr Rahbar
The simulation results were compared to a standard
amount of fitness is. In this diagram the horizontal
axis and the vertical axis represents the number of
generations of genetic algorithm is the fitness values.
Comparing the Best Value for the symptoms ο red,
medium blue with pluses and worst of green Light,
fitness values over generations in the chart are
completed.
Figure 5:. Charts the best, average and worst fitness values
based on a generational advance
Charts the best, average and worst fitness values
based on generational progression downward based on
a function of cycle time and number of cars waiting
for the red lights.
mutation_rate = 0.2;
crossover_rate = 0.5;
So the simulation fitness of the best values to the
next generation of 85 is achieved in the second half
generation algorithm has converged to a solution. In
the second half of the difference between the best and
average generation is observed that due to the
relatively high rate of mutation is a mutation.
According to my tests, this figure is considered one of
the best solutions, the values of the basic parameters
related to the shape of the table below. Other
comparison criteria of simulation-based cycle time
and number of cars waiting to be lowest. Since the
normalization values are based on values between
zero and one is shown in Fig.
The normal values for the variables x1 to x14 after
applying the genetic algorithm is as follows: These
values are arranged from top to bottom as descending.
Due to the number of duplicate values are rounded to
two decimal places are. Since the end of the next
generation of genetic algorithms converge to the
optimum solution is therefore indicates duplicate
values for parameters relevant achieve the best results.
V. COMPARED WITH RELATED WORKS:
Figure 6: Fitness based on increased fitness
Figure 6 on the best, average and worst fitness
values[25], and Figure 7 are based on the best fitness
values. In both figures the horizontal axis represents
the number of generations of the GA. The result is
then compared to the project after 85 generations the
best solution convergence is done. But the article
again after 10,000 generations, better solutions may
also be found.
Figure 7: Comparison of actual and simulated data of the project
www.ijorcs.org

Figure 8: Comparison of actual and simulated data in article 2
The horizontal axis in Figure 8, the numbers of
parameters x1 to x14 is and the vertical axis represents
the difference seated normalized values obtained from
simulation and actual values is even in the worst case
[26], the difference between 0.77 percent and 23
percent error. The difference is 1.00, the best value
reaches zero. In Figure 4-13, each of the numbers on
the horizontal axis, such as AT, BT and ... represent
actual values and the values obtained from the
simulations. The vertical axis is the percent difference
between these two values. Each vertical bar represents
the number 10 is the cycle time for traffic lights.
Results: In the worst case, the project is 23% of actual
flow and the difference in average and worst case
scenarios the difference is very small, almost zero.
The aim of the project is due to algorithms that better
results were reached.
VI. CONCLUSION
The main problem in optimal control traffic control
traffic lights at the crossroads, or in many other
places. Intersections to regulate traffic flow of
vehicles and eliminate conflicting traffic flows are
used, are important elements of an urban traffic
network. As bad as traffic lights control, causing
severe traffic congestion. So one of the most
challenging issues in the field of traffic lights
controlling the traffic management is. Optimal timing
of urban intersection traffic lights with intelligent and
advanced installation and reduced delay, queue length,
Stop vehicles, resulting in reduced fuel consumption
and mitigate the effects of pollution and
environmental degradation due to excessive stopping
vehicles is a traffic light intersection.
Modeling and simulation are widely used in industry.
In fact, the modeling and simulation of an industrial
system is studied before creating economically and
when it is affordable. The aim of this thesis was to
provide a smart way to control traffic. With the
objective of collecting statistical data, a data set
obtained by the genetic algorithm optimization is the
next step. The optimal output of Petri nets CPN
TOOLS modeling and software have been
implemented. It is known that other data collected
crossings and Evolutionary techniques such as
Genetic algorithms are applied to reduce the waiting
time of the traffic lights are red lights and determine
appropriate cycle. Rates of mutation and crossover
operators in genetic algorithm should be such that the
initial generation of diversity in the population
increases the high mutation rate and the convergence
rate are needed to end generations of crossing rate is
high. In this project, two criteria were used to compare
results. First measure the fitness levels of members of
the different generations of the genetic algorithm and
the second criterion is the optimal value of each
parameter is introduced in the coding stage. Due to the
convergence of genetic algorithm to the optimal
solution after 85 generations and no change can be
concluded that ate parameters like mutation and
crossover rate as well as the fitness function was
performed correctly. The graphs are presented and
compared to the results of this project and other
articles based on objective data, Chart based on
minimum levels of fitness as well as the actual data
and the data obtained from the simulations indicate
that this is the approach taken in this project would
achieve better results. Benefits: Petri net modeling
tool is suitable for simulating the traffic lights at
intersections. Colored Petri Nets is a type of a variable
when it is used in this project as well; leading to
accurate simulation of traffic is close to reality. Petri
nets together to study the fractionation system
feasible. Application of Petri Net tool is so broad that
it is designed in Matlab software. Petri net-based
control systems using the event resulting in the
stabilization and synchronization can occur based on
our review. The example shows that the Petri Net is a
simple and convenient tool for modeling complex
systems, which is very useful for the study of
concurrency and uncertainty. Using the model of Petri
nets can be provided on the results of several
intelligent algorithms to be evaluated. Evolutionary
algorithms such as Genetic selection of the best
answers will win instead of only one answer.
Disadvantages: The problem of finding a set of
statistical data so that the data set as an objective of
the project is compiled. Obtaining the proper fitness
www.ijorcs.org

18 Dadmehr Rahbar
function of trial and error is required to find the best
solutions so that it is necessary to adjust the
parameters of the problem as a chromosome and
fitness function in a specific way they fit. Suggestions
Since genetic algorithms may get stuck in local
optimum answers happens it can be combined with
Simulated Annealing algorithm to obtain the best
results. Combination of Genetic Algorithm and
Simulated Annealing gradual hardening stay together
because of their good characteristics, it seems logical
operations. Way to combine the two is used; consider
the appropriateness of mutation rates in genetic
algorithm and the parameter setting of an evolutionary
algorithm is cooling temperature. The combination of
these two algorithms can be such that the temperature
is initialized based on the population and the number
of generations, the next steps are repeated. A couple
of choices aimed at parents and mutation operators are
applied on the cut. Aimed to replace it with a
requirement to produce two children under the age of
each child-parent with the worst fitness value will be
substituted. The combination of Soft Computing
techniques to other evolutionary methods and
techniques of artificial intelligence strategy is suitable
for such problems. Depending on the type of problem
that was addressed in this project, one of the
appropriate algorithms for solving multi-agent
algorithms that due to large number of lights in each
section of a city controlled by separate agents will
eventually is combined optimization methods. All of
these projects can be proposed algorithm and other
algorithms proposed in the implementation of the
CPN Tools, but it would require too much time.
Therefore, it is suggested that the results of statistical
data optimized algorithm in CPN Tools software used
in this project as the work is done.
VII. REFERENCES
[1] H. Zhonghe, C. Yangzhou, S. Jianjun, W. Xu, and G.
Jizhen, “Consensus based Approach to the Signal
Control of Urban Traffic Networks,” Procedia - Soc.
Behav. Sci., vol. 96, no. Cictp, pp. 2511–2522, Nov.
2013.
[2] M. Madireddy, B. De Coensel, A. Can, B. Degraeuwe,
B. Beusen, I. De Vlieger, and D. Botteldooren,
“Assessment of the impact of speed limit reduction and
traffic signal coordination on vehicle emissions using
an integrated approach,” Transp. Res. Part D Transp.
Environ., vol. 16, no. 7, pp. 504–508, Oct. 2011.
[3] D. McKenney and T. White, “Distributed and adaptive
traffic signal control within a realistic traffic
simulation,” Eng. Appl. Artif. Intell., vol. 26, no. 1, pp.
574–583, Jan. 2013.
[4] M. Dotoli, M. P. Fanti, and C. Meloni, “A signal
timing plan formulation for urban traffic control,”
Control Eng. Pract., vol. 14, no. 11, pp. 1297–1311,
Nov. 2006.
[5] S. Barzegar, M. Davoudpour, M. R. Meybodi, a.
Sadeghian, and M. Tirandazian, “Formalized learning
automata with adaptive fuzzy coloured Petri net; an
application specific to managing traffic signals,” Sci.
Iran., vol. 18, no. 3, pp. 554–565, Jun. 2011.
[6] F. Basile, P. Chiacchio, and D. Teta, “A hybrid model
for real time simulation of urban traffic,” Control Eng.
Pract., vol. 20, no. 2, pp. 123–137, Feb. 2012.
[7] F. Corman, A. D’Ariano, I. a. Hansen, and D.
Pacciarelli, “Optimal multi-class rescheduling of
railway traffic,” J. Rail Transp. Plan. Manag., vol. 1,
no. 1, pp. 14–24, Nov. 2011.
[8] H. Mu, J. Yu, and L. Liu, “Evacuation Routes
Optimization with Effects of Traffic Light at
Intersections,” J. Transp. Syst. Eng. Inf. Technol., vol.
11, no. 3, pp. 76–82, Jun. 2011.
[9] H. Dezani, R. D. S. Bassi, N. Marranghello, L. Gomes,
F. Damiani, and I. Nunes da Silva, “Optimizing urban
traffic flow using Genetic Algorithm with Petri net
analysis as fitness function,” Neurocomputing, vol.
124, pp. 162–167, Jan. 2014.
[10] A. Di Febbraro and N. Sacco, “On Evaluating Traffic
Lights Performance Sensitivity via Hybrid Systems
Models,” Procedia - Soc. Behav. Sci., vol. 111, pp.
272–281, Feb. 2014.
[11] F. C. Fang, W. L. Xu, K. C. Lin, F. Alam, and J.
Potgieter, “Matsuoka Neuronal Oscillator for Traffic
Signal Control Using Agent-based Simulation,”
Procedia Comput. Sci., vol. 19, no. Ant, pp. 389–395,
Jan. 2013.
[12] F. Kaakai, S. Hayat, and A. El Moudni, “A hybrid Petri
nets-based simulation model for evaluating the design
of railway transit stations,” Simul. Model. Pract.
Theory, vol. 15, no. 8, pp. 935–969, Sep. 2007.
[13] A. L. C. Bazzan, D. de Oliveira, and B. C. da Silva,
“Learning in groups of traffic signals,” Eng. Appl.
Artif. Intell., vol. 23, no. 4, pp. 560–568, Jun. 2010.
[14] M. P. Fanti, a. Giua, and C. Seatzu, “Monitor design
for colored Petri nets: An application to deadlock
www.ijorcs.org

prevention in railway networks,” Control Eng. Pract.,
vol. 14, no. 10, pp. 1231–1247, Oct. 2006.
[15] T. Becher, “A New Procedure to Determine a User-
oriented Level of Service of Traffic Light Controlled
Crossroads,” Procedia - Soc. Behav. Sci., vol. 16, pp.
515–525, Jan. 2011.
[16] Z.-J. Ding, X.-Y. Sun, and B.-H. Wang, “Violating
traffic light behavior in the Biham-Middleton-Levine
traffic flow model,” Procedia Eng., vol. 31, no. 2011,
pp. 1072–1076, Jan. 2012.
[17] K. Dahal, K. Almejalli, and M. A. Hossain, “Decision
support for coordinated road traffic control actions,”
Decis. Support Syst., vol. 54, no. 2, pp. 962–975, Jan.
2013.
[18] D. Formanowicz, A. Sackmann, P. Formanowicz, and
J. Błazewicz, “Petri net based model of the body iron
homeostasis.,” J. Biomed. Inform., vol. 40, no. 5, pp.
476–85, Oct. 2007.
[19] J. Blazewicz, D. Formanowicz, P. Formanowicz, A.
Sackmann, and M. Sajkowski, “Modeling the process
of human body iron homeostasis using a variant of
timed Petri nets,” Discret. Appl. Math., vol. 157, no.
10, pp. 2221–2231, May 2009.
[20] D. Rahbari, “High Performance Data mining by
Genetic Neural Network”, International Journal of
Computer Science and Business Informatics, ISSN:
1694-2108 | Vol. 5, No. 1. SEPTEMBER 2013.
[21] D. Rahbari, “Hybrid Evolutionary Game Theory in
QoS Routing of Wireless Mesh Networks”,
International Journal of Computer Science and
Telecommunications, Volume 4, Issue 9, September
2013.
[22] D. Rahbari, “A Novel approach in Classification by
Evolutionary Neural Networks”, International Journal
of Mechatronics, Electrical and Computer Technology,
Vol. 4(10), Jan, 2014, pp. 33-52, ISSN: 2305-0543,
P33-52.
[23] D. Rahbari, “ Digital handwritten recognition by
optimized neural networks”, International Journal of
Computers & Technology (IJCT), Volume 13 No 9, pp
35-43.
[24] D. Rahbari, “ A Novel approach in Classification by
Evolutionary Neural Networks”, International Journal
of Mechatronics, Electrical and Computer Technology,
Vol. 4(10), Jan, 2014, pp. 33-52, ISSN: 2305-0543.
[25] J. S. Medina, M. G. Moreno, N. A. D. Ugarte and
E.R.Royo, “Simulation times Vs. Network Size in a
Genetic Algorithm Based Urban Traffic Optimization
Architecture”.
[26] R. A. Ganiyu, S. O. Olabiyisi, E .O . Omidiora, O.
Okediran and O. O. Alo, Modelling and simulation of a
multi-phase traffic light controlled T-type junction
using timed colored petri nets, AMERICAN
JOURNAL OF SCIENTIFIC AND INDUSTRIAL
RESEARCH© 2011, Science Huβ,
http://www.scihub.org/AJSIR. ISSN: 2153-649X
doi:10.5251/ajsir.2011.2.3.428.437.
www.ijorcs.org

Help the Genetic Algorithm to Minimize the Urban Traffic on Intersections

More Related Content

What's hot (20)

Similar to Help the Genetic Algorithm to Minimize the Urban Traffic on Intersections (20)

More from IJORCS (20)

Recently uploaded (20)

Help the Genetic Algorithm to Minimize the Urban Traffic on Intersections