International Journal on Recent and Innovation Trends in Computing and Communication

ISSN: 2321-8169
686 692

Volume: 1 Issue: 9

______________________________________________________________________

AUTOMATED TIMETABLE GENERATOR USING

PARTICLE SWARM OPTIMIZATION
Andeep Kumar
Dept. of Computer Science
Punjabi University, Patiala
Punjab, India
[email protected]

Kawaljeet Singh
University Computer Centre
Punjabi University, Patiala
Punjab, India
[email protected]

Neeraj Sharma
Dept. of Computer Science
Punjabi University, Patiala
Punjab, India
[email protected]

Abstract: The timetabling problem at universities is an

NP-hard problem under multiple constraints and limited
resources. Thus a technique that can handle constraints is
needed to optimize the problem. This paper focuses on
Particle Swarm Optimization (PSO) for finding optimal
solutions to the problem of course Timetabling at a Punjabi
University Patiala. PSO is a promising scheme for solving
NP - hard problems due to its fast convergence and fewer
parameter settings. There are two objectives in this. First
provide a detailed introduction to the topic of timetabling,
Particle Swarm Optimization their method and their
variations. The second objective is to apply them to the
problem of Course Timetabling. The proposed algorithm is
tested using the timetabling data from Department of
Computer Science, Punjabi University, Patiala.

set of desirable conditions for classes and teachers, should

be satisfied.

Keywords: Particle Swarm Optimization (PSO), university

timetabling

5. Some classes need to be held consecutively. For

example the Labs.

1. Introduction: Timetabling relates all activities with

regards to making a timetable. Wren in 1996 defines
timetabling as the allocation, subject to constraints, of
given resources to objects being placed in space time, in
such a way as to satisfy as nearly as possible a set of
desirable objectives. As a result, a timetable specifies which
people will meet at which location and at what time. A
timetable must meet a number of requirements and should
satisfy the desires of all people as well as possible. At
universities, there are many different courses, so there i s no
conflict free timetable available for every student within
that given time. Therefore faculty tries to find the timetable
with the least conflicts. [1]

6. Classrooms must be large enough to hold the

class scheduled in it.

2. Problem Description: The construction of automated

course timetables for academic institutions is a very
difficult problem with a lot of constraints that have to be
respected and a huge search space to be explored, even if
the size of the problem input is not significantly large, due
to the exponential number of the possible feasible
timetables. On the other hand, the problem itself does not
have a widely approved definition, since different
departments face different variations of it. This problem has
therefore proven to be a very complex. Timetables are
considered feasible provided the so-called hard constraints
are respected. However, to obtain high-quality timetabling
solutions, soft constraints, which impose satisfaction of a

2.1 Hard Constraints [5]: Hard constraints are the

constraints that physically cannot be violated; a timetable in
presence of violation of such hard constraints can never be
acceptable. Following are the list of hard constraints:
1. Classrooms must not be double booked.
2. Every subject must be scheduled exactly once in
a day.
3. Lecturers must not be double booked.
4. A lecturer must not be booked when he/she is
unavailable.

2.2 Soft constraints [5]: Some constraints are less straight

forward to define. Usually, these constraints must be fulfilled
as well as possible. The timetable that violates these
constraints is still usable, but it is not convenient for either
students or teachers. Following are the soft constraints:
1. Teachers may prefer specific time slots.
2. Teachers may prefer specific rooms.
3. Some lecturers do not wish to have classes assigned
consecutively in time.
4. Most students and some lecturers do not wish to have
empty periods in their timetables.
5. Classes should be distributed evenly over the week.
6. Classrooms should not be booked which are much
larger than the size of the class.
2.3 The Proposed Mathematical Programming
Model[6]
In order to study the computational effort involved in
solving the problem of interest, the following mathematical
programming model is proposed.
We define the following sets to be used in the model:
I
set of all teachers
686

IJRITCC | September 2013, Available @ http://www.ijritcc.org

______________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

ISSN: 2321-8169
686 692

Volume: 1 Issue: 9

______________________________________________________________________
J
K
L
M
C

set of all courses

set of all subjects
set of all days available
set of all time periods available
number of classrooms available per timeperiod

Finally, the following decision variables will be requiredto

define the problem:
Xijklm=1if teacher I teaches course j subject k on day l and
at time period m; 0 otherwise (iI, jJ, kKj ,lL, mM)
Xmi = Sum of no. of class rooms that is allocated to all
teachers at particular slot m, 1 if teacher I teaches at timeslot
m; 0 otherwise (mM,iI)
Pik= Lies between 1 and 3, each teacher teaches at least one
subject and at most three subjects(iI,kK)
Li=load of teacher I per week(iI)
For our problem, the objective function reflects a
preference function that needs to be maximized. It refers to
the total preferences of assigning courses to the teachers.
The objective function is described by the expression in
equation (1):
Maximize

(1)

The following depicts some of the main constraints

encountered in our timetabling problem
1 (j J, l L, m M)
(2)

most one course section in a particular time period.

(i I, m

M)

(5)

Equation (5) represents the constraint that at each time

period, the number of course sections taught by teachers
could not be more than the number of classrooms available.
(i I, l L)
(6)
Equation (6) calculates the load of each teacher per week.
3. Particle swarm optimization
Kennedy and Eberhart developed Particle swarm
optimization algorithm in 1995. It is a technique based on
particles, each particle has velocity in search space in order
to find feasible solution and adjust its position according its
own previous experience and neighbors experiences. The
initial position and velocity of each particle is anomaly
determined. When each particle to move a new position then
it will remember its personal best (Pbest) and own
information. Each particle will also exchange its information
to other particles and remember its global best (Gbest). Each
particle has fitness value. Fitness value of each particles
current position is compared with Pbest and Gbest. If it is
better than Pbest and Gbest then update the Pbest and Gbest.
Equations (1,2) as shown below are applied to update the
position and velocity of each particle. Fig 3.1 show the flow
chart of standard PSO.[4]
Vid=Vid + c1 * r1 * (Pid - Xid) + c2 * r2 *(Pgd - Xid) (1)
Xid=Xid+Vid

(2)

In the equation

Equation (2) ensures that for a particular course, only one

or zero subject conducted in every time period.
(i (3)
1
3

Vid is the velocity component of ith particle in the dth

dimension

Equation (3) represents the minimum and maximum

number of subjects taught by each teacher. It is assumed
that each teacher has to teach at least one course and at
most three subjects.
1 (i I, l L, m M) (4)

c1 and c2 are learning factors

Equation (4) ensures that each teacher can only teach at

r1 and r2 are random numbers between [0,1]

Xid is the position component of ith particle in the dth

dimension

Pid is the position component of the Pbest of ith

particle in dth dimension
Pgd is the position component of the Gbest in dth
dimension

687
IJRITCC | September 2013, Available @ http://www.ijritcc.org

______________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

ISSN: 2321-8169
686 692

Volume: 1 Issue: 9

______________________________________________________________________
4. Department Timetabling Problem
The four main factors in course timetabling are teachers,
courses, timeslots and classrooms together with other
teaching facilities. The combination of these four factors is
defined as the particle position and each particle represents
a solution group. The objective is to obtain the optimal
particle position.
First we set the four parameters of the PSO that are
number of iterations, learning factors (cognitive and social)
and inertia weight factor. We take the Gbest values for
different number of iterations for 10, 20 ,, 240, 250,
different values of learning factors and different values of
inertia weights. For which value of these parameters, Gbest
is maximum that value is given optimal solution. We set
these parameters and use dataset in our algorithm. We are
generating the automatic timetable for Department of
Computer Centre, Punjabi University Patiala. Thus we take
the dataset from department and use this in our technique.
[7]
The evolving process isas follows and then shows the
department-timetabling flowchart.
(1)

The faculty gives the preferences of subjects and

software labs that they want to teach.
(2) Formulate the linear mathematical model for course
scheduling on the basis of preferences.
(3) The system assigns load to faculty based
on designation.
(4) The system sorted the preference of each faculty
member on the basis of his or herweight.
(5) Set the parameters and implement the particle swarm
algorithm for optimization results.
(6) Particle Swarm optimization allocates the subjects and
software labs to faculty and calculate the objective
function.

Fig 3.1Standard PSO flowchart

3.1 The Inertia Weight: Shi and Eberhart (1998) proposed
the inertial weight value concept and added an inertial
weight value(w) to the original PSO algorithm. The inertia
weight value is used to balance the global search ability and
the local search ability, as shown in Equation (3,4) and
thereby boost the capability to locate the optimal solution
and convergence rate. We use this technique in our paper.
Vid = wVid + c1 * r1 * (Pid - Xid) + c2 * r2 *(Pgd - Xid)
(3)
Xid = Xid + Vid
(4)
w is Inertia weight value
688
IJRITCC | September 2013, Available @ http://www.ijritcc.org

______________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

ISSN: 2321-8169
686 692

Volume: 1 Issue: 9

______________________________________________________________________
5. Results: This study was conducted by analyzing a
situation involving 14 lecturers, 7 classrooms, 28 subjects,
12 software labs and 5 classes. The goal was to produce the
most satisfactory class schedule to meet the various
constraints as well as the expectations of the teachers. In
order to demonstrate the efficiency and performance of the
proposed PSO algorithm, several experiment results were
carried out.
5.1 Iteration Quantity Iteration:
Figure 5.1 shows the results of analysis of the evolution of
10, 20, 30 ., 240, 250 iterations on the given data set
and find the Global best (fitness) value of course
timetabling problem. When the evolution of each iteration
of particles is conducted, the best Fitness value comes from
when the number of iterations is 140.
5.2 Learning Factor Analysis
Using the better average satisfaction value obtained,
analyzing learning factor c1 and c2, testing results with
PSO are shown in figure 5.2 and 5.3.The average fitness
value for c1= 0.1is higher than the other values (figure 5.2).
Same as the average fitness value for c2= 1.8 is higher than
others (figure 5.3).
5.3 Inertia Weight Analysis:
The inertia weight value is used to balance the global
search ability and the local search ability. If the w value
setting is inappropriate, it is impossible to explore unknown
areas and the search for the personal best and global best
will also be affected. Figure 5.4 shows the satisfaction
value obtained from tests conducted with PSO using
different w values. Apparently, there is higher satisfaction
obtained when w= 0.9.

Fig 5.1Iterations verse Gbest (faculty=14,c1=0.12, c2=1.2,w=0.9)

689
IJRITCC | September 2013, Available @ http://www.ijritcc.org

______________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

ISSN: 2321-8169
686 692

Volume: 1 Issue: 9

______________________________________________________________________

Fig 5.2 c1 verse Gbest (faculty 14, iterations 140, c2=1.2, w=0.9)

Fig 5.3c2 verse Gbest (faculty 14, iterations 140, c1=0.1,w=0.9)

Fig 5.4w verse Gbest (faculty 14, iterations 140, c1=0.1,c2= 1.8)
Finally, course timetabling that achieves the optimal
satisfaction for the dataset from department of computer
science is simulated using PSO when iterations=140,
c1=0.9, c2=1.8 and w=0.9. We take twenty eight subjects
(S1, S2, ., S27, S28), twelve labs (L29, L30,..,
L40) and fourteen teachers (T1, T2. T14) in our
paper. T1 to T6 are professors, T7 is associate professor

and T8 to T14 are assistant professors. Each professor can

teach at most one theory subject and one software lab, each
associate professor can teach at most two theory subjects
and one software lab and each assistant professor can teach
at most three theory subjects and one software lab. Each
faculty member fills their preferences for which subject
they want to teach. Every faculty member fills three theory
690

IJRITCC | September 2013, Available @ http://www.ijritcc.org

______________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

ISSN: 2321-8169
686 692

Volume: 1 Issue: 9

______________________________________________________________________
subjects and two software labs. They give a weight for
every preferred subject 6 for most preferred subject, 5 for
average and so onand 1 for least preferred subject.
Similarly 3 for most preferred software lab, 2 for average

and 1 for least preferred software labs. Table 5.1 shows the
preferences for subjects given by teachers and table 5.2
shows the lab preferences and table 5.3 shows the subjects
and
labs
that
allocate
to
teachers.

5.4 Subject preferences by teachers

S1
S2
S3
S4
S5
S6
S7
S8
S9
S10
S11
S12
S13
S14
S15
S16
S17
S18
S19
S20
S21
S22
S23
S24
S25
S26
S27
S28

T1
2
0
0
3
0
0
0
0
0
0
0
0
4
0
0
0
0
0
5
0
0
1
0
0
0
6
0
0

T2
0
0
2
0
0
1
0
3
0
0
0
0
0
0
0
4
0
0
0
0
0
0
0
5
0
0
0
6

T3
T4
T5
T6
T7
T8
T9
0
0
3
0
0
0
2
1
0
0
1
0
0
0
0
0
0
0
5
0
0
0
0
4
0
0
1
4
0
1
0
0
0
0
0
0
0
0
0
0
5
0
0
0
0
6
0
0
0
0
0
0
0
1
0
3
0
0
0
0
3
0
0
0
0
5
0
0
4
0
0
0
0
0
0
0
0
6
0
0
0
6
0
5
0
2
0
0
0
0
0
0
3
0
2
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6
0
0
6
0
3
4
0
0
0
0
0
0
0
0
4
0
0
0
0
0
0
0
0
0
1
5
0
0
0
0
0
0
0
0
0
0
0
3
0
0
6
0
5
4
0
0
0
0
0
0
0
0
6
4
0
2
0
2
0
0
0
0
0
3
0
0
0
0
5
0
0
0
2
0
0
0
1
0
0
0
0
Table 5.1 subject preferences by teachers

T10
0
0
2
0
0
0
1
0
0
0
0
0
0
4
0
0
0
5
0
0
0
0
0
0
6
0
0
3

T11
0
5
0
0
3
0
0
0
2
0
0
0
0
0
1
0
0
6
0
0
0
4
0
0
0
0
0
0

T12
2
0
0
0
0
1
0
0
0
0
4
0
0
0
3
0
0
0
0
5
0
0
0
0
6
0
0
0

T13
0
0
0
5
0
0
0
3
0
0
0
6
0
0
0
4
0
0
0
0
0
0
0
0
2
0
1
0

T14
3
0
0
0
0
0
6
0
4
0
0
0
5
0
0
0
0
0
0
0
0
2
0
0
0
1
0
0

5.5 Labs preferences by teachers

L29
L30
L31
L32
L33
L34
L35
L36
L37
L38
L39
L40

T1
1
0
0
0
2
0
0
0
3
0
0
0

T2
0
0
1
0
0
0
2
0
0
0
3
0

T3
T4
T5
T6
T7
T8
0
0
0
1
0
2
0
0
1
0
0
0
0
3
0
0
0
0
0
0
0
2
0
0
1
0
0
3
0
0
0
2
0
0
2
0
0
0
0
0
0
1
0
0
2
0
0
0
3
0
0
0
3
0
2
0
0
2
0
0
0
0
3
0
1
3
0
1
0
0
0
0
Table 5.2 lab preferences by teachers

T9
0
0
2
0
1
0
0
0
0
0
0
3

T10
0
2
0
0
0
3
0
0
1
0
0
0

T11
0
0
3
0
0
0
2
0
0
0
1
0

T12
2
0
0
0
0
0
3
0
0
0
0
1

T13
0
2
0
0
3
0
0
1
0
0
0
0

T14
1
0
0
2
0
0
0
0
0
3
0
0

5.6 Subject allocation to teachers

691
IJRITCC | September 2013, Available @ http://www.ijritcc.org

______________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

ISSN: 2321-8169
686 692

Volume: 1 Issue: 9

______________________________________________________________________
We implement this dataset in our system using particle
swarm optimization and subjects and

software labs are allocated to teachers as shown below:

TEACHER ID
ALLOCATED SUBJECTS AND SOFTWARE LABS ID
S26,P37
T1
P39,S28
T2
S21,P38
T3
S27,P31
T4
S17,P36
T5
S7,S19,P33
T6
S3,S12,S23
T7
S6,S10,P29
T8
S24,S4,P40
T9
S25,S14,P34
T10
S22,S18,S2,P35
T11
S20,S11,S15
T12
S4,S16,S8,P30
T13
S13,S9,S1,P32
T14
Table 5.3 subjects and software labs allocated to teachers
After the allocation of subjects to teachers system calculate
the objective function. The objective function should be
maximized. As much as objective function large the results
would be more optimize. When we run our system the
calculated objective function is 173. This value gives more
satisfaction results. From our results we see linear model
try to satisfy the all faculty members not only professors.
We observe that junior teachers more satisfy using this
technique.

Particle Swarm Optimization for University Timetabling

Problem, ISSR 1(1)2009,1-21
[6] Swarm Optimization With Local Search, Algorithms
(2013), 227-24
[7] Elizabeth Montero, Maria- Cristina Riff, Leopodlo
Altamirano, A PSO Algorithm to Solve a Real Course +
Exam Timetabling Problem,ICSI (2011), 241-248.

6. Conclusion: PSO is promoting technique for solving

complex problem such as course timetabling problem. Thus
this work discusses the application of inertia weight factor
a type of PSO to find solutions to solve university course
timetabling. Concurrently, to reduce the computational
complexity, particle encoding is designed on the basis of
timeslot. The solutions found in accordance with the
characteristics of the problem have been able to improve
the satisfaction of the teachers and classes toward the
schedule. Any conflicts between the teacher schedules, the
class schedules, or the classroom schedules are also in this
work.
References:
[1] Ioannis X. Tassopoulos, Grigorios N. Beligiannis, A
hybrid particle swarm optimization based algorithm for
high school timetabling problems, in applied soft
computing 12(2012), 3472-3489
[2] Johor Bahru, Safaai, D. ; Zaiton, M., A Combination of
PSO and Local Search in University Course Timetabling
Problem, IEEE(2009),492-495
[3]Liu Su-hua, Li Lin, Study of course scheduling based on
particle swarm optimization , IEEE(2011), 1692-1695
[4]Shu-Chuan Chu, Yi-Tin Chen ; Jiun-Huei Ho, Timetable
Scheduling Using Particle Swarm Optimization,
IEEE(2006)
[5] Ho Sheau Fen Irene, Safaai Deris, Siti Zaiton Mohd
Hashim, Incorporating of Constraint-Based Reasoning into
692
IJRITCC | September 2013, Available @ http://www.ijritcc.org

______________________________________________________________________