Hindawi

Advances in Fuzzy Systems

Volume 2018, Article ID 1274969, 19 pages
https://doi.org/10.1155/2018/1274969

Research Article
Fuzzy Dynamic Parameter Adaptation in ACO and PSO for
Designing Fuzzy Controllers: The Cases of Water Level and
Temperature Control

Fevrier Valdez ,1 Juan Carlos Vazquez,1 and Fernando Gaxiola 2

1
Tijuana Institute of Technology, Tijuana, BC, Mexico
2
Autonomous University of Chihuahua, Chihuahua, CHIH, Mexico

Correspondence should be addressed to Fevrier Valdez; [email protected]

Received 19 January 2018; Revised 23 March 2018; Accepted 22 April 2018; Published 2 July 2018

Academic Editor: Roberto Sepúlveda

Copyright © 2018 Fevrier Valdez 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.

A novel approach applied to Particle Swarm Optimization (PSO) and Ant Colony Optimization is presented. The main contribution
of this work is the use of fuzzy systems to dynamically update the parameters for the ACO and PSO algorithms. In the case
of ACO, two fuzzy systems are designed for the Ant Colony System (ACS) algorithm variant. The first system adjusts the value
for the pheromone evaporation parameter from the global pheromone trail update equation and the second system adjusts the
values for the pheromone evaporation parameter from the local pheromone trail update equation. In the case of PSO, a fuzzy
system is designed to find the values for the inertia weight parameter from the velocity equation. Fuzzy logic controllers (FLCs)
are optimized with ACO and PSO, respectively, to prove the performance of the proposed approach. The particular benchmark
problems considered to test the proposed methods are the water level control in a tank and temperature control in a shower.
Therefore, PSO and ACO algorithms are applied in the optimization of the parameters of the FLCs. The achievement of the proposed
fuzzy ACO and PSO algorithms is compared with the original results of each benchmark control problem.

1. Introduction mediated by the environment. ACO algorithms are utilized

to solve many paradigmatic problems, like the Travelling
In automating the design of fuzzy controllers, the use of Salesman Problem (TSP) [5, 6]. Different ACO variants have
metaheuristic algorithms has been amply proposed. The been used for fuzzy system controller problems models [7, 8].
proposed approach is based on the swarm intelligence models Those algorithms have focused on fuzzy logic parameter
[1] that perform research of the collective behavior in decen- tuning to find the values to each membership function,
tralized schemes; two swarm intelligence models are utilized the membership functions type, the number of membership
in this paper: the Ant Colony Optimization (ACO) [2] and functions, and so forth [9, 10].
Particle Swarm Optimization (PSO) [3, 4] algorithms. The The inspiration of PSO is based on the social behavior of
development of the swarm intelligence models is founded on flock of birds or fishes. The procedure of PSO to encounter
imitating the social behavior of living beings, particularly of the optimal solution is based on the task of this animal
insects or animals, in performing the search for the optimal community. In a PSO method, in the search space, a swarm
solution in a solution space for a problem. of particles (individuals) is scattered. Each particle can be an
ACO is a metaheuristic algorithm based on the collabo- aspirant solution for the problem at optimization. The PSO
ration of artificial ants in a colony to search for an optimal has been implemented, obtaining optimal results, to different
solution for complex problems. In ACO algorithms, the optimization problems, such as benchmark functions [11, 12],
collaboration between the artificial ants is an important part, finding optimal results for permutation problems [13, 14], and
and this consists in assigning the computational resources to optimizing the training of neural networks [15, 16] and we
artificial ants (artificial agents) by indirect communication also found optimization of fuzzy inference systems [17] and
2 Advances in Fuzzy Systems

Input x
(x)
(y) y
x
Fuzzifier Inference Plant
Defuzzifier
Engine

output
States or
Inference
Engine

Figure 1: Scheme of fuzzy logic control (FLC).

Low Medium High Table 1: Fuzzy rules for 𝑐ℎ𝑤, 𝑐1 , and 𝑐2 .

1
0.9 Input Output
Degree of Membership

0.8 Rule number

0.7
𝑁𝑈 𝑤 𝑐ℎ𝑤 𝑐1 𝑐2
0.6 1 1 1 3 1 1
0.5 2 1 2 1 3 3
0.4
3 1 3 1 3 3
0.3
0.2 4 2 1 3 1 1
0.1 5 2 2 2 1 3
0 6 2 3 1 3 3
1 2 3 4 5 6 7 8 9 10
7 3 1 3 1 1
Figure 2: Membership functions for the inputs and outputs of the
8 3 2 1 3 3
fuzzy inference system.
9 3 3 1 3 3

c1(3) Table 2: Fuzz rules 𝜌.

Input Output
Rule number
NU (3) Δ𝜏𝑏𝑠 𝜏 𝜌
c2(3) 1 1 1 3
2 1 2 2
3 1 3 1
W (3) c3(3) 4 2 1 3
5 2 2 2
Figure 3: Fuzzy inference system to adjust the cℎ𝑤, 𝑐1 , and 𝑐2 6 2 3 1
parameters. 7 3 1 3
8 3 2 2
9 3 3 1
some optimization problems for fuzzy control systems [18],
even with other hybrid optimization algorithms [19].
Unlike other algorithms that have been applied for con- pheromone evaporation parameter in the equations used to
trol problems, we have proposed the design of fuzzy systems calculate the update of the local and global pheromone. For
to adjust some parameters for each of the optimization the PSO algorithm, we have designed a fuzzy inference system
algorithms. The algorithms, both ACO and PSO, are applied to adjust dynamically in each iteration the parameter of the
to optimize control systems [20, 21]. For this paper, we have inertia weight in the equation used to obtain the velocity of
considered two benchmark control problems for designing the particle. The approaches were applied to the optimization
their fuzzy controllers with the proposed approach [22]. of water level control and temperature control in a shower for
The novelty in the contribution of this work is the testing.
development of fuzzy systems to dynamically adjust param- In the PSO parameters control problems in this work,
eters of the ACO and PSO algorithms [23, 24]. In the the membership functions will help in automating the fuzzy
ACO algorithm, we have proposed the use of two fuzzy control [25, 26]. The rules were framed through numerous
inference systems to dynamically adjust in each iteration the simulations, which are carried out to determine the best
Advances in Fuzzy Systems 3

tb-s (3)

p (3)

t (3)
Figure 6: Water tank representation.
Figure 4: Fuzzy system to adjust the pheromone evaporation 𝜌.

Level (3)
p (3)

Valve (3)
(3)
Rate (3)
t (3)
Figure 7: Original representation of the fuzzy system for water tank.
Figure 5: Fuzzy system to adjust the pheromone evaporation 𝜉.

Table 3: Fuzz rules 𝜉.

possible set of rules aimed at pushing the stability limits of
the system to its maximum. The membership functions can Input Output
Rule number
be estimated by studying the behavior of different conditions 𝑝 𝜏 𝜉
and for different contingency cases [27, 28]. 1 1 1 3
The organization of this paper is as follows: Section 2 2 1 2 2
describes some basic concepts of fuzzy logic, Section 3 3 1 3 3
provides a brief description of the fuzzy logic controller
4 2 1 2
(FLC), and Section 3 explains the standard PSO and ACO
5 2 2 2
algorithms with the basic structure for water level control in a
shower in a tank and the structure for temperature control in 6 2 3 2
a shower and the experimental results made in this research. 7 3 1 1
The conclusions are summarized in Section 4. 8 3 2 2
9 3 3 3
2. Background
2.1. Fuzzy Logic Control (FLC). The fuzzy logic control is
utilized in ill-defined complex process that can be operated problem [3, 4]; the process consists of flying around a search
by a trained human without knowing the dynamics of the space, and the position of each particle is updated using the
system. personal experience and the social particles experience [29,
In a FLC, the basic idea consists in utilizing the knowledge 30]. The position of each particle is updated by the following
of an expert operator for the construction of the FLC that equation:
performs the control for a system; the input-output variables
for the system are represented by fuzzy rules (IF-THEN)
𝑥𝑖 (𝑡 + 1) = 𝑥𝑖 (𝑡) + V𝑖 (𝑡 + 1) (1)
at difference to a complicated mathematical model. In the
design of the FLC, the use of fuzzy rules with linguistic
variables and the fuzzy reasoning allows incorporating the where 𝑥𝑖 (𝑡 + 1) describes the position of the particle 𝑖 in
experience of the human expert. A traditional scheme for a the time (𝑡 + 1), 𝑥𝑖 (𝑡) describes the position of the particle 𝑖 in
FLC is shown in Figure 1. the time (𝑡), and V𝑖 (𝑡 + 1) denotes the velocity of the particle
𝑖 in the time (𝑡 + 1).
2.2. Particle Swarm Optimization. PSO works with particles In this work, we utilized the gbest PSO variant, in which
for representing a potential solution to the optimization the experience neighborhood for each particle is obtained
4 Advances in Fuzzy Systems

from the entire swarm. For each particle 𝑖, the velocity is

calculated with the equation as follows:

V𝑖𝑗 (𝑡 + 1) = V𝑖𝑗 (𝑡) + 𝑐1 𝑟1𝑗 (𝑡) ⌊𝑦𝑖𝑗 (𝑡) − 𝑥𝑖𝑗 (𝑡)⌋ Temp (3) Cold (5)
(2)
+ 𝑐2 𝑟2𝑗 (𝑡) ⌊𝑦̂𝑗 (𝑡) − 𝑥𝑖𝑗 (𝑡)⌋

where V𝑖𝑗 (𝑡) describes the velocity of the particle 𝑖 in the

dimension 𝑗 in the time 𝑡, 𝑥𝑖𝑗 (𝑡) represents the position of the
particle in the time 𝑡, 𝑦𝑖𝑗 (𝑡) denotes the best position of the Flow (3) Flow (5)
particle, 𝑦̂𝑗 (𝑡) describes the global best position of all parti-
Figure 8: Original representation of the fuzzy system for tempera-
cles, 𝑐1 , 𝑐2 are cognitive and social components, respectively
ture control.
(positive acceleration constants), and 𝑟1𝑗 (𝑡), 𝑟2𝑗 (𝑡) are random
numbers in the interval [0, 1].
Some research work has been developed to enhance the
To obtain the new 𝑐ℎ𝑤, 𝑐1 , and 𝑐2 values, nine fuzzy
convergence and the optimal solution obtained with the basic
rules are used in the fuzzy inference system. An
PSO, like the velocity constriction [31], inertia weight [32],
example of one fuzzy rule is given as follows:
and velocity clamping [27].
In this paper, a modification of the gbest PSO variant If 𝑁𝑈 is MEDIUM, and
using the inertia weight (𝑤) is proposed. The inertia weight
allows having control in the exploration and exploitation 𝑤 is HIGH
swarm besides the control of the speed and direction of the then 𝑐1 is HIGH, 𝑐2 is HIGH and 𝑐ℎ𝑤 is LOW
particles. Equation (2) of the gbest PSO changes as follows:
The rules were obtained by empirical knowledge obtained
V𝑖𝑗 (𝑡 + 1) = 𝑤V𝑖𝑗 (𝑡) + 𝑐1 𝑟1𝑗 (𝑡) ⌊𝑦𝑖𝑗 (𝑡) − 𝑥𝑖𝑗 (𝑡)⌋ by experimenting and are shown in Table 1.
(3)
+ 𝑐2 𝑟2𝑗 (𝑡) ⌊𝑦̂𝑗 (𝑡) − 𝑥𝑖𝑗 (𝑡)⌋ The range of 𝑁𝑈 is in [1, 10], the value for 𝑤 is in 0.2 ≤
𝑤 ≤ 1.2, and the values of 𝑐1 and 𝑐2 are in 1.0 ≤ 𝑐1 ,
A number of works to dynamically adjust the inertia 𝑐2 ≤ 2.0.
weight 𝑤 have been developed, like the random adjustments The mathematical descriptions for 𝑐1 and 𝑐2 are
[33], linear decreasing [26, 34], nonlinear decreasing [34], defined with the following expressions:
and fuzzy adaptive inertia [35]. We propose adjusting the 𝑟
𝑐1 𝑐
inertia weight 𝑤 and learning factors 𝑐1 and 𝑐2 of the PSO ∑𝑖=1 𝜇𝑖 1 (𝑐1𝑖 )
during the iteration process. 𝑐1 = 𝑟𝑐1 𝑐
(4)
∑𝑖=1 𝜇𝑖 1
The learning factors 𝑐1 and 𝑐2 affect the total velocity of
a particle [36]. The cognitive component (𝑐1 ) determines the
own confidence of the particle and the social component (𝑐2 ) where 𝑐1 is percent of cognitive acceleration of the
defines the confidence of the particle in the neighbors. particle 𝑖, 𝑟𝑐1 is number of fuzzy rules activated to
𝑐
We have designed a fuzzy inference system for the 𝑐1 , 𝑐1𝑖 is output of the fuzzy rule 𝑖 for 𝑐1 , and 𝜇𝑖 1 is
adaptation of the inertia weight and learning factors; the membership function value of fuzzy rule 𝑖 for 𝑐1 .
design consists of the following: 𝑟
𝑐2 𝑐
∑𝑖=1 𝜇𝑖 2 (𝑐2𝑖 )
Two inputs: the number of iterations (𝑁𝑈) when the 𝑐2 = 𝑟𝑐2 𝑐
(5)
∑𝑖=1 𝜇𝑖 2
best fitness does not change and the actual value of the
inertia weight (𝑤)
where 𝑐2 is percent of cognitive acceleration of the
Three outputs: the change in inertia weight (𝑐ℎ𝑤) and particle 𝑖, 𝑟𝑐2 is number of fuzzy rules activated to
the change in learning factors 𝑐1 and 𝑐2 𝑐
𝑐2 , 𝑐2𝑖 is output of the fuzzy rule 𝑖 for 𝑐2 , and 𝜇𝑖 2 is
Three membership functions for each input and membership function value of fuzzy rule 𝑖 for 𝑐2 .
outputs in the fuzzy inference systems: they are shown
in Figure 2, defined as LOW, MEDIUM, and HIGH
and represented as a left triangle, middle triangle, and For obtaining the values of the variable, the Mamdani
right triangle, respectively FIS model is utilized.
Figure 3 illustrates the fuzzy inference system architecture The method of the centroid of area is performed to
with two inputs (the number of iterations when the best obtain the defuzzification:
fitness does not change (𝑁𝑈) and the actual inertia weight
∫𝑧 𝑀𝑜 (𝑧) 𝑧𝑑𝑧
(𝑤)) and three outputs (the change in inertia weight (𝑐ℎ𝑤) 𝑂= (6)
and learning factors (𝑐1 ) and (𝑐2 )). ∫𝑧 𝑀𝑜 (𝑧) 𝑑𝑧
Advances in Fuzzy Systems 5

Flow

soft good hard

Degree of Membership

Particle

1. Trapezoidal.
2. Triangular.
3. Gaussian.
4. Bell.
Type of
membership function: Parameter values

Figure 9: Example of the graphical representation of the PSO particles for the membership functions.

3. Ant Colony Optimization added. This action is called pheromone evaporation and is
performed as follows:
The first algorithm to fall into the framework of the ACO
metaheuristics was the Ant System (AS) [2]. In AS, 𝑘 (arti- 𝜏𝑖𝑗 ←󳨀 (1 − 𝜌) 𝜏𝑖𝑗 (8)
ficial) ants build, at the same time, a graph for the solution
of the optimization problem. At first, the nodes for the ants where 0 < 𝜌 ≤ 1 is the pheromone evaporation rate.
are randomly chosen. At each building step, to obtain the The parameter 𝜌 is utilized to elude the accumulation without
node in the next iteration of the ant 𝑘, a probabilistic action limits of the pheromone trails and it allows the algorithm to
selection rule, called random proportional rule, is applied “forget” wrong decisions previously taken. After applying the
[37]. Determining the probability of the ant 𝑘 to select the evaporation, the ants add pheromones on the arcs of their
path from node 𝑖 to 𝑗 is as follows: route as follows:
𝑚
𝛼
[𝜏𝑖𝑗 ] [𝜂𝑖𝑗 ]
𝛽 𝜏𝑖𝑗 ←󳨀 𝜏𝑖𝑗 + ∑ Δ𝜏𝑖𝑗𝑘 (9)
𝑝𝑖𝑗𝑘 = 𝛼 𝛽
(7) 𝑘=1
∑𝑙∈𝑁𝑖𝑘 [𝜏𝑖𝑗 ] [𝜂𝑖𝑗 ]
where Δ𝜏𝑖𝑗𝑘 is the quantity of pheromones that the ant 𝑘
provides on the arcs visited. It is determined as follows:
𝜂𝑖𝑗 defines the heuristic value that is independent of 1
{ 𝑘 , arc (𝑖, 𝑗) ∈ 𝑇𝑘
the experience of the ant. Δ𝜏𝑖𝑗𝑘 = {𝑐 (10)
𝛼,𝛽 are two parameters that define the influence of the {0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
pheromone trail and the heuristic information.
𝑁𝑖𝑘 represents the probable neighborhood of the ant 𝑘 where 𝑐𝑘 represents the longitude of the graph and 𝑇𝑘
in the node 𝑖, namely, the grouping of nodes that the is calculated as the sum of the longitudes of the arcs
ant 𝑘 has not visited yet. of the route for the 𝑘th ant.
𝜏𝑖𝑗 defines the pheromone evaporation for the route.
Many research works dealing with the updating rules for
The update of the pheromone trails is performed after Δ𝜏𝑖𝑗𝑘 have been performed, like the Ant System (AS) [2], Elitist
all the ants build their routes. To reach the update, first the Ant System (EAS) [38], Rank-Based Ant System (ASrank )
pheromone values on the arcs are reduced by a constant factor [39], Max-Min ant System (MMAS) [40], and Ant Colony
and later, on the arcs which the ants passed, pheromone is System (ACS) [41].
6 Advances in Fuzzy Systems

FOR each particle i

FOR each dimension d
Initialize position xid randomly
Initialize velocity vid randomly
END FOR
END FOR
Iteration k=1
DO
FOR each particle i
Calculate fitness value
IF the fitness value is better than p bestid in history
Assign current fitness value as the p bestid
END IF
IF the fitness value is better than g bestid in history
Assign current fitness value as the g bestid
END IF
END FOR
FOR each particle i
FOR each dimension d
Calculate w, c1 and c2 with the fuzzy inference system
Calculate velocity using the equation
vid (k+1) = w vid (k) + c1 Rand1 (pid - xid ) + c2 Rand2 (𝑝𝑔𝑑 - xid )
Update particle position using the equation
xid (k+1) = xid (k) + vid (k+1)
END FOR
END FOR
k=k+1
WHILE Maximum iterations or mı́nimum error criteria are not attained

Algorithm 1: PSO algorithm.

FOR each ant i

Initialize ant i
END FOR
Iteration k=1
DO
FOR each particle i
Construct the trail
Select the next node
Calculate the value with the fuzzy inference system
Update the local pheromone trail
IF the trail is complete
Calculate the value with the fuzzy inference system
Update the global pheromone trail
ELSE
GO to Select the next node
END IF
END FOR
k=k+1
WHILE Maximum iterations or mı́nimum error criteria are not attained

Algorithm 2: ACO algorithm.

In this paper, we used an Ant Colony System (ACS) [41] deposit and the pheromone evaporation only in the arcs of
for three important reasons. At first instance, the ACS uses the more promising route. Third, when the ant moves from
an action choice rule that is more aggressive than the AS node 𝑖 to node 𝑗 using the arc (𝑖, 𝑗), with the objective of
in the exploitation of the search experience concentrated by incrementing the exploration, it eliminates a small amount
the strongest ants. Second, the ACS applies the pheromone of pheromone from the arc [42, 43].
Advances in Fuzzy Systems 7

Start

Define Parameters

Create Particles
(Swarm)

No
i<NumParticles

Yes
Calculate Fitness

No Fuzzy Controller
Fitness<Pbest

Yes
Pbest=swarm

No
Pbest<Gbest

Yes

Pbest=Gbest

No
i<NumParticles

Yes

Update Velocity (4)

Proposed Fuzzy
Update Position (2) System

No Stopping Criteria

Yes
Stop

Figure 10: The flowchart of PSO.

ACO has a process in which it does not use heuristic { argmax , 𝑖𝑓 𝑞 ≤ 𝑞0

values and the part of the consequent can be merely selected Δ𝜏𝑖𝑗𝑘 = {𝑗∈𝑁𝑖𝑘 [𝜏𝑖𝑗 ]𝛼 (12)
as follows: {𝐽 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝛼
[𝜏𝑖𝑗 ]
𝑝𝑖𝑗𝑘 = (11) where 𝑞 is random value uniformly distributed in
𝛼
∑𝑙∈𝑁𝑖𝑘 [𝜏𝑖𝑗 ] [0, 1], 𝑞0 is parameter in the interval [0, 1], and 𝐽 is
random value calculated by (11) (with 𝛼 = 1).
In ACS, the ant 𝑘 moves from node 𝑖 to node 𝑗 according In the global pheromone trial update for ACS, in each
to a pseudo-random-proportional rule as follows: iteration, only the global best ant has authorization to add the
8 Advances in Fuzzy Systems

input output

Level Rate Valve

PARTICLES
T= Type of membership function
V= Parameter Value

Figure 11: Representation of the PSO for the optimization of membership functions.

input output
Flow Temp Cold Hot

PARTICLES

T= Type of membership function

V= Parameter Value

Figure 12: Representation of PSO for the optimization of membership functions.

Three membership functions for each input and

output: LOW, MEDIUM, and HIGH, respectively,
represented as membership functions of left triangle,
middle triangle, and right triangle. Figure 4 illustrates
the fuzzy system to obtain the change in pheromone
evaporation 𝜌
Nine fuzzy rules are considered from which the
changes in 𝜌 are calculated. The rules are obtained by
experimental knowledge as shown in Table 2
The ranges of Δ𝜏𝑏𝑠 and 𝜏 are normalized into [0, 1]
In ACS, besides the global pheromone trail update rule, a
Figure 13: Variable division and ant search path. local pheromone update rule is applied after the ants cross an
arc (𝑖, 𝑗) in the construction of the graph. This rule is given by
the following equation:
pheromone. Accordingly, the update in ACS is performed as
𝜏𝑖𝑗 ←󳨀 (1 − 𝜉) 𝜏𝑖𝑗 + 𝜉𝜏0 (14)
follows:
where 𝜉, 0 < 𝜉 < 1. The value of 𝜏0 is configured to take
𝜏𝑖𝑗 ←󳨀 (1 − 𝜌) 𝜏𝑖𝑗 + 𝜌Δ𝜏𝑖𝑗𝑏𝑠 (13) the initial value of the pheromone trails.
We have defined a fuzzy inference system for pheromone
where Δ𝜏𝑖𝑗𝑏𝑠 = 1/𝐶𝑏𝑠 ; in ACS, the pheromone trail update evaporation 𝜉 which consists of the following components:
is applied to the arcs of 𝜏𝑏𝑠 , not to all the arcs as in AS. Two inputs: the maximum probability of choosing the
We have defined a fuzzy inference system for pheromone
next node 𝑝𝑖𝑗𝑘 and the pheromone trail 𝜏𝑖𝑗
evaporation 𝜌 which is made up of the following elements:
One output: the change in pheromone evaporation 𝜉
Two inputs: the value of Δ𝜏𝑖𝑗𝑏𝑠 and the pheromone trail Three membership functions for each input and
𝜏𝑖𝑗 output: LOW, MEDIUM, and HIGH, respectively,
One output: the change in pheromone evaporation 𝜌 represented as membership functions of left triangle,
Advances in Fuzzy Systems 9

Start

Define Parameters

Create Ants (Colony)

No
j<NumAnts

Yes
Tour Construction

Fuzzy
Find next node (12) Controller

Local Pheromone
Trail Update (15)
Proposed Fuzzy
No System
Tour is finished

Yes
Global Pheromone
Trail Update (14)
Proposed Fuzzy
System

Stopping criteria
No
Yes
Stop

Figure 14: The flowchart of the ACO algorithm.

input output

level rate valve

Type of
membership function: 1 1 1 1 1 1 1 1 1 1 1
2
1. Trapezoidal.
2. Triangular.
3. Gaussian.
4. Bell. 4 4 4 4 4 4 4 4 4 4 4
Type of membership function
input
output
level rate valve
-1 -1 -1 -1 -0.1 -0.1 -0.1 -0.1 -1 -1 -1 -1 -1 -1 -1 -1
-0.9 -0.09

1 1 1 1 0.1 0.1 0.1 0.1 1 1 1 1 1 1 1 1

Parameter values

Figure 15: Representation of the potential solutions in the pheromone matrix for the water system.
10 Advances in Fuzzy Systems

input output

flow temp cold hot

Type of
1 1 1 1 1
membership function: 1 1 1 1 1 1 1 1 1 1 1

2
1. Trapezoidal.
2. Triangular.
3. Gaussian.
4. Bell.
4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

Type of membership function

input
output
flow temp cold hot
-20 -20 -20 -20 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1

-19 -0.9

20 20 20 20 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Parameter values

Figure 16: Representation of the potential solutions in the pheromone matrix for temperature system.

Input Output
Level
1 high okay low
Degree of membership

0.8
0.6 Valve
1 closefast closeslow nochange openslow openfast
0.4
Degree of membership

0.8
0.2
0.6
0
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1 0.4
Rate 0.2
1 negative none positive
0
Degree of membership

0.8 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1

0.6
0.4
0.2
0
−0.1 −0.08 −0.06 −0.04 −0.02 0 0.02 0.04 0.06 0.08 0.1

Figure 17: Membership functions for the water tank FLC without optimization.

middle triangle, and right triangle. Figure 5 illustrates [22]. For this case, the control of the water that flows into the
the fuzzy inference system to obtain the change in tank is made by using one valve. The outflow rate depends
pheromone evaporation 𝜉 on the diameter of the output pipe, which is constant, and
Nine fuzzy rules are considered from which the the pressure in the tank, which varies with the water level.
changes in 𝜉 are calculated. The rules are obtained by Therefore, the system has nonlinear characteristics. The main
experimental knowledge as shown in Table 3 objective of the fuzzy logic controller (FLC) for the water
The ranges of 𝑝 and 𝜏 are normalized into [0, 1] tank is to maintain the water in between the lower level L
and upper level U according to the reference R, as shown
3.1. Water Level Control in a Tank. The water tank model and in Figure 6. The FLC performs the control of the level and
the control of this water tank using a fuzzy control system is velocity to the water, giving as output of the FLC the quickness
the system used for the experiments of the proposed approach of flow out of the pipe.
Advances in Fuzzy Systems 11

Simulation (4) If (level is OKAY) and (rate is POSITIVE) then (valve

1.8 Reference
Control signal in CLOSE SLOW)
1.6
(5) If (level is OKAY) and (rate is NEGATIVE) then
1.4
(valve in OPEN SLOW)
Reference

1.2
1 3.2. Temperature Control in a Shower. We also used as basis
0.8 for our experiments the problem of temperature control and
0.6 the fuzzy system presented in [22]. As Figure 8 illustrates, the
fuzzy controller consists of two inputs: the water temperature
0.4
0 1000 2000 3000 4000 5000 6000 7000 and the flow velocity. The controller utilizes the outputs
Time obtained with the two inputs to control the cold and hot
valves.
Figure 18: Behavior of the water level in the tank with the FLC not
optimized. For the fuzzy system of the temperature control, 16 mem-
bership functions were used: 3 per each input and 5 at the
output and all this is illustrated in Figure 8. The fuzzy variable
1.8 Simulation temp is defined to have three fuzzy sets, COLD, GOOD, and
Reference
1.6
Control signal HOT, with associated membership functions as leftTriangle,
Triangle, and rigthTriangle, respectively; the fuzzy variable
1.4
flow is defined to have three fuzzy sets, SOFT, GOOD, and
1.2 HARD, with associated membership function as leftTriangle,
1 Triangle, and rigthTriangle, respectively; the fuzzy variables
cold and hot are defined to have five fuzzy sets, CLOSEFAST,
0.8
CLOSESLOW, STEADY, OPENSLOW, and OPENFAST, with
0.6 associated membership functions as leftTriangle, Triangle,
0.4 and rigthTriangle, respectively. The following nine fuzzy rules
0 2000 4000 6000 8000 10000 12000 were used to calculate the change of the flow and temperature
Time in the outputs:
Figure 19: Behavior for the optimized water tank FLC with PSO.
(1) If (temp is COLD) and (flow is SOFT) then (cold is
OPENSLOW) and (hot is OPENFAST)
The controller needs to be efficient to manage flow in the (2) If (temp is COLD) and (flow is GOOD) then (cold is
pipe and maintain the current level of water in the tank. The CLOSESLOW) and (hot is OPENSLOW)
water level error is the input for the controller (the difference (3) If (temp is COLD) and (flow is HARD) then (cold is
between the desired water level and the current water level) CLOSEFAST) and (hot is CLOSESLOW)
and the rate of opening or closing for the valve is the output. (4) If (temp is GOOD) and (flow is SOFT) then (cold is
As Figure 7 illustrates, the fuzzy controller has two inputs: OPENSLOW) and (hot is OPENSLOW)
the rate of change of the water level and the difference
between the actual and the required water levels. The con- (5) If (temp is GOOD) and (flow is GOOD) then (cold is
troller uses these inputs to manipulate the inflow rate. STEADY) and (hot is STEADY)
For the fuzzy system of the water tank, 11 membership (6) If (temp is GOOD) and (flow is HARD) then (cold is
functions were used: 3 per each input and 5 at the out- CLOSESLOW) and (hot is CLOSESLOW)
put and all this is illustrated in Figure 7. The linguistic (7) If (temp is HOT) and (flow is SOFT) then (cold is
variable level is defined to have three fuzzy sets, HIGH, OPENFAST) and (hot is OPENSLOW)
OK, and LOW, with the associated membership functions
as leftGaussian, Gaussian, and rightGaussian, respectively; (8) If (temp is HOT) and (flow is GOOD) then (cold is
the linguistic variable rate is defined to have three fuzzy OPENSLOW) and (hot is CLOSESLOW)
sets, NEGATIVE, NONE, and POSITIVE, with associ- (9) If (temp is HOT) and (flow is HARD) then (cold is
ated membership functions as leftGaussian, Gaussian, and CLOSESLOW) and (hot is CLOSEFAST)
rigthGaussian, respectively; the linguistic variable valve is
defined to have five fuzzy sets, CLOSE FAST, CLOSE SLOW, 3.2.1. Implementation of the Proposed Methodology. In this
NO CHANGE, OPEN SLOW, and OPEN FAST, with asso- section, the PSO and ACO algorithms are implemented
ciated membership function as leftTriangle, Triangle, and for the optimization of two FLC systems [44, 45]. The
rigthTriangle, respectively. The following five fuzzy rules are optimization approach has been designed in the following
used to calculate the change in the valve: sequence: first, the optimization defines the type of mem-
(1) If (level is OKAY) then (valve in NO CHANGE) bership function: can be Gaussian, triangular, trapezoidal,
or generalized bell; later, the optimization determines the
(2) If (level is LOW) then (valve in OPEN FAST) parameters for the chosen membership functions; it is for
(3) If (level is HIGH) then (valve in CLOSE FAST) the interesting values depending on the type of membership
12 Advances in Fuzzy Systems

Input Output
Level
high okay low
1

0.8
Degree of membership

0.6 Valve
closefast closeslow nochange openslow openfast
1
0.4
0.8

Degree of membership
0.2
0.6
0

−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1 0.4

Rate 0.2
negative none positive
1
0
0.8
Degree of membership

−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1

0.6

0.4

0.2

−0.1 −0.08 −0.06 −0.04 −0.02 0 0.02 0.04 0.06 0.08 0.1

Figure 20: Membership functions optimized with PSO for the water tank FLC.

Simulation
Reference
3.3. Proposed Methodology for PSO. In this case, we rep-
2
Control signal resented the problem as is shown in Figure 9, where one
particle can be divided into sections to represent the type and
1.5
parameters of the membership function.
Reference

PSO is used to optimize fuzzy controllers for two cases.

1
In one, PSO is used to optimize the FLC from the water
0.5
control in a tank and in the other PSO is used to optimize
the temperature control in a shower. They both used a set
0
of 200 particles and 300 iterations and they perform 30
0 2000 4000 6000 8000 10000 12000 14000 16000 18000 experiments. We use 200 particles and 300 iterations because
Time after making several tests the best results were obtained with
Figure 21: Behavior for the optimized water tank FLC with ACO. these values. In Section 3.6.3, we show a statistical test with
the 30 experiments against ACO to validate in the best way
the experimental results. The flowchart and pseudocode of
function, like the coordinates of the points of edge, the the PSO algorithm are displayed in Figure 10 and Algorithm 1
standard deviation, and so forth. of PSO, respectively, and the process is described as follows:
The objective function is built to represent the fitness of a
(1) In the search space of the problem, we use random
solution and is considered as an interface between optimiza-
values for the positions and velocities of the particles;
tion problem and the algorithm. In order to determine the
this is to create and initialize the population.
fitness of the fuzzy model, the mean square error (MSE) is
used (defined in (15)). The goal error for MSE is zero. (2) The fitness (in the fuzzy controller) is evaluated for
all particles using the established objective function,
1 𝑛 2 MSE.
𝑀𝑆𝐸 = ∑ [𝑦 (𝑘) − 𝑦̃ (𝑘)] (15)
𝑁 𝑘=1
(3) A comparison between the current fitness against
the previous fitness of each particle is applied. If the
̃ is calculated output
where 𝑦(𝑘) is desired output, 𝑦(𝑘) current fitness is better than the previous fitness, then
of the proposed methodology, and N is number of the current value is established for the pbest and the
pieces of data by the validation of model. current position like the pbest position.
Advances in Fuzzy Systems 13

Input Output
Level
okay
1 high low
Degree of membership

0.8

0.6 Valve
closefast closeslow nochange openslow openfast
1
0.4
0.8

Degree of membership
0.2
0.6
0

−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1 0.4

Rate
0.2
negative none positive
1
0
0.8
Degree of membership

−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1

0.6

0.4

0.2

−0.1 −0.08 −0.06 −0.04 −0.02 0 0.02 0.04 0.06 0.08 0.1

Figure 22: Membership functions optimized for the water tank FLC with ACO.

Input Output

Temp Cold
cold good hot closefast closeslow steady openslow openfast
1 1
Degree of membership

Degree of membership

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

−20 −15 −10 −5 0 5 10 15 20 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1
Flow Hot
soft good hard closefast closeslow steady openslow openfast
1 1
Degree of membership

Degree of membership

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1

Figure 23: Membership functions not optimized for the temperature control FLC.
14 Advances in Fuzzy Systems

1 Simulation
Reference
particles, where we can see that each subswarm is formed
Control signal
0.9 with two value types such as type of membership function
0.8 and parameter value corresponding to each variable (input-
0.7 output).
Reference

0.6
0.5 3.4. Proposed Methodology for ACO. In the Ant Colony
0.4 Optimization algorithm, the problem to optimize must be
0.3 represented as a combinatorial problem; in this case, the types
0.2 of membership functions should be represented as a graph; it
0 200 400 600 800 1000 1200 1400 1600 means that the potential values require to be represented as
Time a graph. This is an important part, for the reason that ACO
Figure 24: Behavior of the not optimized temperature control in the operates with this type of representation. Each node must be
FLC for the shower. a possible solution that an ant can take.
The ant can consider each node like a potential solution
Simulation
to choose. At the end of the iterations, every ant will have a
1 Control signal
Reference
graph with the potential values to make the evaluation; in this
graph, if we suppose that the problem to optimize is used to
0.8 design a graph with qi nodes, the node value is determined
Reference

as 𝑥𝑖𝑗 with i as the variable index (i=1, 2,. . ., n) and j as the

0.6
number node for xi (j = 1, 2, qij ). After the division, a network
0.4 that is composed of variables and their division is constructed
as depicted in Figure 13.
0.2 The flowchart for ACO is displayed in Figure 14, and the
0 0.5 1 1.5 2 2.5
Time x 104
illustration of the ACO algorithm is shown in Algorithm 2
and the process is described as follows:
Figure 25: Behavior for the optimized temperature control FLC
with PSO. (1) The ant colony is created and initialized with random
positions. We set the initial parameters for ACO
making several trial and error experiments. For each
(4) A comparison among the current fitness of each ant, construct a graph (for the fuzzy controller).
particle against the best global fitness (gbest) of the
previous population is applied. If the current fitness (1.1) The next node is selected in function to the
is better than gbest, then the index and value of the probability obtained with (11).
current particle are established like the gbest. (1.2) Reduce the local pheromone trail (evaporation)
(5) The evaluation is performed with the star topology: according to (14) with the proposed fuzzy sys-
tem (Figure 7).
(5.1) In base to (3) and the proposed fuzzy inference
system (Figure 3), the particle velocity is calcu- (2) Reduce the global pheromone trail (evaporation)
lated. according to (14) with the proposed fuzzy system
(Figure 6).
(5.2) Actualize the particle position using (1).
(3) Steps (2) to (4) are repeated until the algorithm
(6) Steps (2) to (6) are repeated until the algorithm reaches the maximum iterations or the best fitness is
reaches the maximum number of iterations or the best better than the desired value.
fitness is better than the desired value.
3.4.1. Water Level Control in a Tank. In this paper, the design
3.3.1. PSO for Water Level Control. Figure 11 shows the of fuzzy logic control (FLC) is performed using Ant Colony
graphical representation of the particles used for the opti- System. Figure 15 shows the graphical representation of the
mization of the FLC for tank control. As it was mentioned potential solutions in the pheromone matrix used for the
before, we have implemented a set of subswarms of size optimization of FLC from tank control, where the first node
of 27 particles, where we can see that each subswarm is of the graph represents the nest and the food source is the
formed with two value types, such as the type of membership last node. This implementation is structured with a total of 27
function and parameter value corresponding to each variable nodes, where each node is formed for a type of membership
(input-output). function or parameter value (input-output).

3.3.2. PSO for Temperature Control. Figure 12 shows the 3.4.2. Temperature Control in a Shower. We also used the
graphical representation of particles used for the optimiza- Ant Colony System algorithm for this case. Figure 16 shows
tion of the FLC from tank control. As mentioned before, the graphical representation of the potential solutions in the
we have implemented a set of subswarms of size of 40 pheromone matrix used for the optimization of FLC from
Advances in Fuzzy Systems 15

Input Output
Temp Cold
1 cold good hot closefast closeslow steady openslow openfast
1

Degree of membership
0.8 0.8
Degree of membership

0.6 0.6

0.4 0.4

0.2 0.2

0 0

−20 −15 −10 −5 0 5 10 15 20 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1
Flow Hot
soft good hard closefast closeslow steady openslow openfast
1 1

Degree of membership
Degree of membership

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1

Figure 26: Membership functions optimized for the temperature control FLC with PSO.

Simulation Control signal optimization, the type of membership functions, and their
1 Reference
structure.
0.8 Figure 18 shows the original simulation in relation with
Reference

the Mean Square Error (described in Section 3) obtaining an

0.6 error of 0.0668 when the fuzzy system was not optimized,
where 𝑦̃ is taken as the control signal, 𝑦 as the reference signal,
0.4
and N as the number of data points.
0.2
0 2000 4000 6000 8000 10000 12000
Time 3.5.2. Simulation Results for Water Level Control with PSO.
The best experiments, with the smaller error of 0.026952, are
Figure 27: Behavior for the optimized temperature control FLC
presented in Figure 19; Figure 19 shows the behavior of the
with ACO.
reference signal against the control signal in simulation for
the optimization with PSO; the resulting new structures of
tank control, where the first node of the graph represents the membership functions after parameter optimization for the
nest and the food source is the last node. This implementation water tank FLC are shown in Figure 20.
is structured with a total of 40 nodes, where each node is
formed for a type of membership function or parameter value 3.5.3. Simulation Results for Water Level Control with ACO.
(input-output). The best experiments, with a smaller error of 0.045792, are
presented in Figure 21; Figure 21 shows the behavior of the
3.5. Simulation Results. In this part, the results of the opti- reference signal against the control signal in simulation for
mization of the FLCs for the water level control and tem- the optimization with ACO; the resulting new structures of
perature control systems with Particle Swarm Optimization membership functions after parameter optimization for the
and Ant Colony Optimization are presented. For comparison water tank FLC are shown in Figure 22.
purposes, we used FLCs provided by [22] which have not
been optimized. 3.6. Simulation Results for Temperature Control. Figure 23
shows the fuzzy logic control (FLC) for the shower tempera-
3.5.1. Simulation Results for the Water Level Control. Figure 17 ture without optimization, the type of membership functions,
shows the fuzzy logic control (FLC) for the water tank without and their structure.
16 Advances in Fuzzy Systems

Input Output
Temp Cold
1 cold good hot closefast closeslow steady openslow openfast
1
Degree of membership

Degree of membership
0.8
0.8
0.6 0.6

0.4 0.4

0.2 0.2

0 0
−20 −15 −10 −5 0 5 10 15 20 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1
Flow Hot
soft good hard closefast closeslow steady openslow openfast
1 1
Degree of membership

Degree of membership
0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1

Figure 28: Membership functions optimized for the temperature control FLC with ACO.

Figure 24 shows the original simulation in relation with Table 4: Student’s 𝑡-test.
the Mean Square Error (described in Section 3) obtaining an 𝑁 Mean Standard deviation SE mean
error of 0.007013 when the fuzzy system was not optimized,
30 0.190 0.163 0.030
where 𝑦̃ is taken as the control signal, 𝑦 as the reference signal,
and N as the number of data points. 30 0.807 0.434 0.079
Difference = 𝜇 (ACO) - 𝜇 (PSO).
Estimate for difference: -0.6164.
3.6.1. Simulation Results for Temperature Control with PSO. 95% lower bound for difference: -0.7593.
The best experiment, with a smaller error of 0.0003758, is 𝑇-value = -7.28
presented in Figure 25; Figure 25 shows the behavior of the 𝑃 value = 1.20587𝐸 − 08.
reference signal against the control signal in simulation for
the optimization with PSO; the resulting new structures of
membership functions after parameter optimization for the results with a Student’s 𝑡-test. We use 30 samples, selected
shower temperature FLC are shown in Figure 26. randomly from the 100 experiments to make this test. We
can see in Table 4 for this case that the ACO method was
3.6.2. Simulation Results for Temperature Control with ACO. better than PSO because the T-value was -7.28; however, in
The best experiment, with a smaller error of 0.00076356, is Table 5, we can see that the 30 experiments and the values
presented in Figure 27; Figure 27 shows the behavior of the are very similar with both methods; however, some values of
reference signal against the control signal in simulation for PSO are larger than 1, and for the ACO algorithm, the values
the optimization with ACO; the resulting new structures of are smaller than 1. Therefore, the differences for these two
membership functions after parameter optimization for the methods using optimized controllers were significant.
shower temperature FLC are shown in Figure 28.
4. Conclusions
3.6.3. Comparative Study of Optimized ACO versus Optimized
PSO: The Case of Temperature Control in a Shower. In this In this work, we have applied the Ant Colony Optimization
section, we included a statistical test to validate in the best way (ACO) and Particle Swarm Optimization (PSO) metaheuris-
our approach between the ACO and PSO optimized methods. tics for the design of fuzzy systems. For this purpose, we
We made 100 experiments for the case of the temperature have adopted one ACO variant, namely, the Ant Colony
control with both methods. The following table shows the System (ACS). Both algorithms, ACS and PSO, have changing
Advances in Fuzzy Systems 17

Table 5: Experimental results with ACO and PSO. with the dynamic parameter adaptation, ACO and PSO can
be good alternatives to solve some cases of control; the
Experiment PSO ACO
experimental results are very similar to the two optimized
1 1.1456 0.0067823 methods; however, for the case of temperature control in a
2 1.2076 0.012086 shower, ACO was slightly superior to PSO. Also, there are
3 0.90314 0.0281 other experiments to validate this approach with other cases,
4 0.42635 0.033326 for example, the water tank control, and in some cases PSO is
5 0.20136 0.051824 better than ACO.
6 0.75873 0.066958 For future work, we plan to hybridize these algorithms
to compare results and test with fuzzy inference systems for
7 0.093008 0.077077
different membership functions or implement type 2 fuzzy
8 0.68108 0.080195 inference systems to enhance the results of the work.
9 1.1802 0.083176
10 1.1365 0.10675
Data Availability
11 1.1563 0.10827
12 0.31743 0.12463 The data used to support the findings of this study are
13 1.2254 0.13095 available from the corresponding author upon request.
14 1.0505 0.13263
15 1.2925 0.13276 Conflicts of Interest
16 0.40378 0.14014
17 1.2477 0.14722
The authors declare that there are no conflicts of interest of
any kind regarding the publication of this paper.
18 1.1119 0.16984
19 0.63035 0.17191
20 1.0559 0.17539 Acknowledgments
21 0.41091 0.22107 The authors would like to express their gratitude to CONA-
22 0.069772 0.22883 CYT and Tijuana Institute of Technology for the facilities and
23 1.2036 0.23622 resources granted for the development of this research.
24 1.1208 0.3004
25 0.078089 0.30398 References
26 1.071 0.30562
27 0.30528 0.4007 [1] J. Kennedy, R. Eberhart, and Y. Shi, Swarm Intelligence, Morgan
Kaufmann Publisher, 2001.
28 1.2911 0.49584
[2] M. Dorigo, V. Maniezzo, and A. Colorni, “Ant system: optimiza-
29 1.1934 0.57976 tion by a colony of cooperating agents,” IEEE Transactions on
30 0.23343 0.65851 Systems, Man, and Cybernetics, Part B: Cybernetics, vol. 26, no.
Error average 0.806756967 0.19036481 1, pp. 29–41, 1996.
[3] J. Kennedy and R. Eberhart, “Particle swarm optimization,”
in Proceedings of the IEEE International Conference on Neural
Networks (ICNN ’95), vol. 4, pp. 1942–1948, Perth, Western
parameters with fuzzy dynamic adaptation during execution. Australia, November-December 1995.
For the ACS algorithm, the global and local evaporation of the
[4] R. C. Eberhart and J. Kennedy, “A new optimizer using particle
pheromone is dynamically adapted using a fuzzy system. For
swarm theory,” in Proceedings of the 6th International Sympo-
the PSO algorithm, the inertia weight and the parameters c1 sium on Micromachine and Human Science, pp. 39–43, Nagoya,
and c2 are adapted dynamically with a fuzzy system. The type Japan, October 1995.
of membership functions and the parameters of membership [5] D. S. Johnson and L. A. McGeoch, “The traveling salesman
function of the fuzzy logic controllers for temperature control problem: a case study,” in Local search in combinatorial opti-
and water level control were optimized with ACO and PSO mization, Wiley-Intersci. Ser. Discrete Math. Optim., pp. 215–
algorithms, respectively. The results of both optimization 310, Wiley, Chichester, 1997.
algorithms are compared. Those results show good solutions [6] G. Reinelt, The Traveling Salesman Problem: Computational
in comparison with the original problem that is not opti- Solutions for TSP Applications, vol. 840 of Lecture Notes in
mized. Computer Science, Springer, Berlin, Germany, 1994.
The use of metaheuristic algorithms is amply utilized in [7] R. Martı́nez-Marroquı́n, O. Castillo, and J. Soria, “Parameter
optimization of problems; for the benchmark problems of tuning of membership functions of a fuzzy logic controller
temperature control and water level control, the implemented for an autonomous wheeled mobile robot using ant colony
fuzzy systems are used to adjust parameters of the ACO and optimization,” in Proceedings of the 2009 IEEE International
PSO and improve the results of the algorithms. Conference on Fuzzy Systems, pp. 2007–2012, kor, August 2009.
As a general conclusion with the statistical test applied [8] E. Lizarraga, O. Castillo, J. Soria, and F. Valdez, “A Fuzzy Control
with both analyzed methods, we can demonstrate that, Design for an Autonomous Mobile Robot Using Ant Colony
18 Advances in Fuzzy Systems

Optimization,” in Recent Advances on Hybrid Approaches for [23] G. K. Venayagamoorthy and S. Doctor, “Navigation of mobile
Designing Intelligent Systems, vol. 547 of Studies in Computa- sensors using PSO and embedded PSO in a fuzzy logic con-
tional Intelligence, pp. 289–304, Springer International Publish- troller,” in Proceedings of the Conference Record of the 2004 IEEE
ing, Cham, 2014. Industry Applications Conference; 39th IAS Annual Meeting, pp.
[9] P. Kaur, Sh. Kumar, and A. Singh, “Optimization of Membership 1200–1206, usa, October 2004.
Functions Based on Ant Colony Algorithm,” International [24] J. Van Ast, R. Babuška, and B. De Schutter, “Fuzzy ant colony
Journal of Computer Science and Information Security, vol. 10, optimization for optimal control,” in Proceedings of the Amer-
2010. ican Control Conference (ACC ’09), pp. 1003–1008, IEEE, St.
[10] H. Jiang, H. Deng, and Y. He, “Determination of fuzzy logic Louis, Mo, USA, June 2009.
membership functions using extended ant colony optimization [25] L. A. Zadeh, “Fuzzy logic,” The Computer Journal, vol. 21, no. 4,
algorithm,” in Proceedings of the 5th International Conference on pp. 83–93, 1988.
Fuzzy Systems and Knowledge Discovery, FSKD 2008, pp. 581– [26] H. Yoshida, Y. Fukuyama, S. Takayama, and Y. Nakanishi, “A
585, chn, October 2008. particle swarm optimization for reactive power and voltage
[11] F. Valdez, P. Melin, and O. Castillo, “Evolutionary method control in electric power systems considering voltage security
combining particle swarm optimization and genetic algorithms assessment,” in Proceedings of the IEEE SMC’99 Conference
using fuzzy logic for decision making,” in Proceedings of the Proceedings. 1999 IEEE International Conference on Systems,
2009 IEEE International Conference on Fuzzy Systems, pp. 2114– Man, and Cybernetics, pp. 497–502, Tokyo, Japan.
2119, kor, August 2009.
[27] Y. Shi and R. Eberhart, “Parameter selection in particle swarm
[12] P. Angeline, “Evolutionary optimization versus particle swarm optimization,” in Evolutionary Programming VII, vol. 1447 of
optimization: philosophy and performance differences,” in Lecture Notes in Computer Science, pp. 591–600, Springer, New
Evolutionary Programming VII, V. W. Porto, N. Saravanan, D. York, NY, USA, 1998.
Waagen, and A. E. Eiben, Eds., vol. 1447 of Lecture Notes in
[28] F. Valdez, P. Melin, and O. Castillo, “Particle swarm optimiza-
Computer Science, pp. 601–610, Springer, Berlin, Germany, 1998.
tion for designing an optimal fuzzy logic controller of a DC
[13] M. F. Tasgetiren, Y. C. Liang, M. Sevkli, and G. A. Gencyil- motor,” in Proceedings of the 2012 Annual Meeting of the North
maz, “A particle swarm optimization algorithm for makespan American Fuzzy Information Processing Society, NAFIPS 2012,
and total flowtime minimization in the permutation flow- usa, August 2012.
shop sequencing problem,” European Journal of Operational
Research, vol. 177, no. 3, pp. 1930–1947, 2007. [29] T. Y. Abdalla, A. A. Abed, and A. A. Ahmed, “Mobile robot
navigation using PSO-optimized fuzzy artificial potential field
[14] J. Salerno, “Using the particle swarm optimization technique to with fuzzy control,” Journal of Intelligent & Fuzzy Systems:
train a recurrent neural model,” in Proceedings of the 9th IEEE Applications in Engineering and Technology, vol. 32, no. 6, pp.
International Conference on Tools with Artificial Intelligence, pp.
3893–3908, 2017.
45–49, Newport Beach, Calif, USA, November 1997.
[30] H. A. Hashim, S. El-Ferik, and M. A. Abido, “A fuzzy logic
[15] P. F. Ribeiro and W. Kyle Schlansker, A Hybrid Particle Swarm
feedback filter design tuned with PSO for adaptive controller,”
and Neuronal Network Approach for Reactive, IEEE, Power
Expert Systems with Applications, vol. 42, no. 23, pp. 9077–9085,
Control, 2006.
2015.
[16] C. R. Eberthart and H. Xiaohui, Human Tremor Analysus Using
[31] M. Clerc, “The Swarm and the Queen: Towards a Deterministic
Particle Swarm Optimization, Purdue School of Engineering
and Adaptive Particle Swarm Optimization,” in Proceedings of
and Technology, Indiana University Purdue University Indi-
the IEEE Congress on Evolutionary Computation, vol. 3, pp. 1951–
anapolis, Indianapolis, 1999.
1957, 1999.
[17] E. E. Omizegba and G. E. Adebayo, “Optimizing fuzzy member-
ship functions using particle swarm algorithm,” in Proceedings [32] Y. Shi and R. Eberhart, “A modified particle swarm opti-
of the 2009 IEEE International Conference on Systems, Man and mizer,” in Proceedings of the IEEE International Conference
Cybernetics, SMC 2009, pp. 3866–3870, usa, October 2009. on Evolutionary Computation and IEEE World Congress on
Computational Intelligence, (Cat. No.98TH8360), pp. 69–73,
[18] D. de la O, O. Castillo, and A. Meléndez, “Optimization of
Anchorage, Alaska, USA, May 1998.
Fuzzy Control Systems for Mobile Robots Based on PSO,” in
Recent Advances on Hybrid Approaches for Designing Intelligent [33] J. Peng, Y. Chen, and R. Eberhart, “Battery pack state of charge
Systems, vol. 547 of Studies in Computational Intelligence, pp. estimator design using computational intelligence approaches,”
191–208, Springer International Publishing, Cham, 2014. in Proceedings of the 15th Annual Battery Conference on Appli-
cations and Advances, pp. 173–177, usa, January 2000.
[19] R. Martı́nez-Soto, O. Castillo, L. T. Aguilar, and A. Rodrı́guez
Dı́az, “A hybrid optimization method with PSO and GA to [34] S. Naka, T. Genji, T. Yura, and Y. Fukuyama, “Practical distribu-
automatically design Type-1 and Type-2 fuzzy logic controllers,” tion state estimation using hybrid particle swarm optimization,”
International Journal of Machine Learning and Cybernetics, vol. in Proceedings of the 2001 IEEE Power Engineering Society
6, no. 2, pp. 175–196, 2015. Winter Meeting, pp. 815–820, usa, February 2001.
[20] L. Amador-Angulo and O. Castillo, “Comparison of the optimal [35] Y. H. Shi and R. C. Eberhart, “Fuzzy Adaptive Particle Swarm
design of fuzzy controllers for the water tank using ant colony Optimization,” in Proceedings of the IEEE Congress on Evolution-
optimization,” Studies in Computational Intelligence, vol. 547, pp. ary Computation, vol. 1, pp. 101–106, IEEE Press, May 2001.
255–273, 2014. [36] R. C. Eberhart and Y. Shi, “Comparing inertia weights and con-
[21] R. Fierro and O. Castillo, “Design of fuzzy control systems with striction factors in particle swarm optimization,” in Proceedings
different PSO variants,” Studies in Computational Intelligence, of the Congress on Evolutionary Computation (CEC ’00), vol. 1 of
vol. 451, pp. 81–88, 2013. (Cat. No.00TH8512), pp. 84–88, July 2000.
[22] R. Goering, “Matlab edges closer to electronic design automa- [37] S.-M. Zhang, D. Huang, S.-C. Chu, T.-W. Sung, and T.-Y.
tion world,” EE Times. Wu, “An adaptive ACO-based node deployment algorithm in
Advances in Fuzzy Systems 19

wireless sensor networks,” Journal of Internet Technology, vol.

18, no. 5, pp. 1193–1202, 2017.
[38] M. Dorigo, V. Maniezzo, and A. Colorni, “Positive feedback
as a search strategy,” Technical report 91-016, Dipartimento di
Elettronica, Politecnico di Milano, Milan, 1991.
[39] B. Bullnheimer, R. F. Hartl, and C. Strauss, “A new rank based
version of the ant system: a computational study,” Central
European Journal of Operations Research, vol. 7, no. 1, pp. 25–38,
1999.
[40] T. Stutzle and H. H. Hoos, “Improving the Ant System: A
detailed report on the MAX-MIN Ant System,” Technical report
AIDA-96-12, FG Intellektik, FB Informatik, TU Darmstadt,
Germany, 1996.
[41] M. Dorigo and L. M. Gambardella, “Ant colony system: a coop-
erative learning approach to the traveling salesman problem,”
IEEE Transactions on Evolutionary Computation, vol. 1, no. 1, pp.
53–66, 1997.
[42] O. Castillo, H. Neyoy, J. Soria, P. Melin, and F. Valdez, “A
new approach for dynamic fuzzy logic parameter tuning in ant
colony optimization and its application in fuzzy control of a
mobile robot,” Applied Soft Computing, vol. 28, pp. 150–159, 2015.
[43] O. Castillo and J. Soria, “Fuzzy Logic for Dynamic Parameter
Tuning in ACO and Its Application in Optimal Fuzzy Logic
Controller Desig,” Fuzzy Logic Augmentation of Nature-Inspired
Optimization Metaheuristics, 2015.
[44] E. Vinodh Kumar, G. S. Raaja, and J. Jerome, “Adaptive PSO for
optimal LQR tracking control of 2 DoF laboratory helicopter,”
Applied Soft Computing, vol. 41, pp. 77–90, 2016.
[45] F. Olivas, F. Valdez, O. Castillo, C. I. Gonzalez, G. Martinez,
and P. Melin, “Ant colony optimization with dynamic param-
eter adaptation based on interval type-2 fuzzy logic systems,”
Applied Soft Computing, vol. 53, pp. 74–87, 2017.
 Advances in
Multimedia
Applied
Computational
Intelligence and Soft
Computing
The Scientific
Engineering
Journal of
Mathematical Problems
Hindawi
World Journal
Hindawi Publishing Corporation
in Engineering
Hindawi Hindawi Hindawi
www.hindawi.com Volume 2018 http://www.hindawi.com
www.hindawi.com Volume 2018
2013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Modelling &  Advances in 

Simulation  Artificial
in Engineering Intelligence
Hindawi Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Advances in

International Journal of
Fuzzy
Reconfigurable Submit your manuscripts at Systems
Computing www.hindawi.com

Hindawi
www.hindawi.com Volume 2018 Hindawi Volume 2018
www.hindawi.com

Journal of

Computer Networks
and Communications
Advances in International Journal of
Scientific Human-Computer Engineering Advances in
Programming
Hindawi
Interaction
Hindawi
Mathematics
Hindawi
Civil Engineering
Hindawi
Hindawi
www.hindawi.com Volume 2018
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

International Journal of

Biomedical Imaging

International Journal of Journal of

Journal of Computer Games Electrical and Computer Computational Intelligence
Robotics
Hindawi
Technology
Hindawi Hindawi
Engineering
Hindawi
and Neuroscience
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018