Customizable algorithm for university timetabling

and scheduling
Amsa Aijaz Vijay Kumar
Hira Ansari
Software Engineering Department Software Engineering Department Software Engineering Department
Mehran university of Engineering and Mehran university of Engineering and Mehran university of Engineering and
technology, Sindh technology, Sindh technology, Sindh
Pakistan Pakistan Pakistan
[email protected] [email protected] [email protected]

Masood Junejo Saeed Shah Muhammad Saad

Software Engineering Department Software Engineering Department Software Engineering Department
Mehran university of Engineering and Mehran university of Engineering and Mehran university of Engineering and
technology, Sindh technology, Sindh technology, Sindh
Pakistan Pakistan Pakistan
[email protected] [email protected] [email protected]

Imran Memon
Zahid Hussain Khaskheli Dr Qasim

Software Engineering Department Department of Computer Science

Software Engineering Department Bahria university, Karachi
Mehran university of Engineering and Mehran university of Engineering and Pakistan
technology, Sindh technology, Sindh [email protected]
Pakistan Pakistan
[email protected] [email protected]

Abstract—University course timetable problem is considered Many additions are introduced in field relating timetabling
to be Np hard problem which increases exponentially with the problems which includes Artificial Intelligence, Operation
size of problem. In this paper we have introduced algorithm
which is made in accordance with Mehran Uet (Pakistan) Time-
Research and computational Intelligence [6], [7] further they
table. Timetabling is a sort of issue in which events (slots, are fragmented into four categories.
teachers, courses, and so on) must be organized into various
slots to such an extent that conflicts must be avoided for given 1. Sequential Method, timetabling problems are
resources. New functions are introduced in accordance with soft considered to be graphical, which uses the timing curves and
and hard constraints to be satisfied accordance with Mehran
Uet Pakistan. assigning them sequentially to valid time slots thus generating
a time table [8].
Keywords— UCTP [17], credit hours ,UET, time tabling
2. Constraint Based Methods, in this timetabling problems are
I. INTRODUCTION (HEADING 1)
taken in concept of assigning events to their resources so that
Problems like university course time table (UCTP[17]) are all constraints needed to be satisfied [9]
considered to be one of the difficult kind as well as commonly
faced by the institutes and universities, they are usually solved 3. Cluster Methods, In this problem is sorted into subsets of
through optimization algorithms in operation research. [1], events each subset's element is described in such a way that it
[2], [3], [4]. Allocation of teachers, students and classrooms meets with expectations of all hard constraints moreover these
to courses. The problem can be defined as the placement of subsets are allocated to real time slots in order to meet all the
resources that are teachers, students and classrooms in a expectations of soft constraints as well[10]
space-time which meets all university(i-e hard and soft
constraints) by optimally utilizing facilities available. 4. Meta heuristic methods, they are mostly derived from
Moreover UCTP[17] becomes hard to be solved by the nature and their approaches to solve a problem are methods
conventional methods because of number of computations taken from nature. [11]
needed which is directly proportional to the problem The proposed algorithm is designed and inspired from Mehran
size.[5].In addition to that UCTP[17] is considered as NP hard Uet time-table which focuses on credit hours and provides
problem which means response time of system grows consistency in slots. In this conducted research it focuses on
exponentially with the size of problem. Going through the generation of the time table with specially designed algorithm,
researches done by generation of time table using Meta which is fast and should be able to generate table in short
heuristic. period of time as the requirements are completed.

978-1-7281-2334-9/19/$31.00 c 2019 IEEE

113
 II. RELATED WORK: This paper proposes a general answer for the School
This paper focuses on a genetic algorithm and a new local timetabling issue. Most heuristic proposed before approaches
Search algorithm fitness function is developed which helps the issue from the students' perspective. This solution, in any
changing the time table feasible according to one's need. case, works from the teachers' perspective for example teacher
Moreover in future the local search can be announced by accessibility for a given schedule vacancy. While all the hard
applying more mega rustic algorithms or increasing the constraints (for example the accessibility of instructors, and so
genetic operators like selection mechanism and crossover time on.) are settled thoroughly, the planning arrangement
and can be tested on different time table data sheets from presented in this paper is a versatile one, with an essential
examination and courses. [12] expect to understand the issue of conflicts of addresses and
subjects, relating to teachers. The intention of the algorithm to
The importance of timetable is to organize the event in a create a time-table schedule automatically is fulfilled. The
composed list. The faculty or college timetable normally algorithm incorporates various systems, meant to enhance the
comprises of colossal information or event with a great deal proficiency of the search operation. It additionally, addresses
of limitation and qualities that unfit to change and should be the vital hard constraint of conflicts between the accessibility
planned. The dependable and fast generation of timetable is of teachers. The non-rigid soft requirements for example
demanding in generating the timetable particularly if the optimizing objectives for the pursuit task are likewise
timetable framework is running in a web base. Unfortunately, adequately handled. Given the generality of the operation, it
there are a great deal of timetable frameworks that unable to can additionally be adjusted to progressively explicit
complete the generation in a brief time-frame. Here, another situations, for example University, examination scheduling
combination of heuristic and random selection was and further be upgraded to make railway time tables.
approached in producing the timetable with quick and reliable Subsequently, through the procedure of automation of the
outcome. The combination was experimented for one time-table issue, numerous hours of making a viable timetable
semester in Faculty of Electrical Engineering of University have been decreased eventually.[15]
Technology Malaysia that comprises of tremendous number
of understudies with parallel addresses, multi-disciplinary Planning timetable is a standout among the most unpredictable
subjects, limited rooms and furthermore lecturer's constraint . and error prone application. There are as yet significant issues
The execution uncovers that this combination is able to like generation of high cost of time while planning and
generate the timetable for the entire faculty in just 16 seconds, scheduling . These issues are repeating as often as possible. In
along these lines, verifying its effectiveness and reliability. this manner there is great requirement for an application
This paper focuses that by utilizing heuristic, limitation based distributing the course equally and without impacts. In this
and arbitrary choice, the outcome demonstrates that it is more paper some mainstream algorithms were analyzed that are
reasonable than micro-GA. Nevertheless, in pre-analysis this utilized for solving this issue. Genetic algorithm, meme
philosophy was not able to foresee whether the generation is algorithm and differential evolution algorithms are the
practical or not based. Therefore in future, a procedure that primary algorithms that were considered, other than it focused
able to overcome this issue ought to be connected for better to comprehend the heuristic approach to solve this problem.
execution. [13]
Execution of three algorithms was analyzed on three factors
For a school, to guarantee its normal high level of teaching speed , time required for tuning, calculation and simplicity.
quality, it must build up a lot of thorough, standardized Genetic algorithm provides the most basic or easy to
teaching plans, and strictly perform them. The course plan and understand solution for the issue it is even extensively quicker
arrangement is the center of the school teaching program, so than both the differential evolution and meme algorithm .But
sensible and precise educational modules plan is fundamental unlike the differential evolution it can't give better outcomes
to enhance the teaching standards. In this paper, through the when provided with noisy data and Differential development
investigation and the synopsis of the current issues and can likewise be connected to wide assortment of issues.
existing problems, a mathematical model for the course Whereas the meme algorithm is by all aspects moderate and
timetable framework is proposed. In the meantime, through requires huge amount of iterations because of the local search
the use of the pattern recognition technology in artificial algorithm embedded in it yet provides progressively
intelligence, aiming at this scientific model another university productive output when contrasted with the Differential
course timetable framework configuration program is evolution and the genetic algorithm. [16]
proposed and figured. This program not exclusively can well
This paper is organized as follows: Section 2 related work ,
explain the deficiencies of the current course timetable
Section 3 describes the methodology. Results and testing are
framework, yet in addition is straightforward and simple to
defined in Section 4. Section 5 concludes the conclusion and
work, has strong versatility. [14]
Future work.

114
 III. METHODOLOGY C. Flowchart:
The algorithm should be designed in such a way that it should
meet expectations of all hard and minimize the violation in
soft constraint as much as possible. In the proposed algorithm
following soft and hard constraints are satisfied.
A. Hard Constraint:
 Each teacher have only one course at a time.
 More than one class can’t be allocate in a single slot.
 No: of classes in a week depends on their credit hours.
 Friday should have five hours and slots should be
accommodate in such a way.
B. Soft Constraints:
 Second time of same course should not assigned in a
same day.
 Practical classes slots must be consecutive, it covers
maximum three slots in a day.
 More than one practical classes cannot be allocated on
a single day.
 Practical classes always come after their theory class.
In this section the proposed system model for designed and
algorithm and its flowchart is given below in figu. re.1
Methodology starts with the assignment of courses which Figure 1: Flowchart of designed algorithm
is further followed by defining the credit hours of course
along with duration of classes, defining all the soft and
hard constraints algorithm then calculates the time slots The flowchart above describes the basic working and flow of
per day it checks whether the constraints are satisfied if it an algorithm used in this study. Furthermore the pseudo codes
meets the requirements it will generates the optimized given on the top adjacent column show the practical working
timetable, if it doesn't then it goes back to defining the time of an algorithm.
slots duration and repeats the process. Pseudo code for calculating slots per day is given as below:

Before proceeding to the making of timetable there are

some pre-requisites that need to be filled.

the pre-requisites are given below:

-- Teachers Data

-- Course Data

-- Department Data

--Student batch and class section's data

--Working days in a week

--Working hours in a week Figure 2 psuedo code 1 for assigning empty timing slots

these factors are to be considered before the scheduling The results for this algorithm is slots per day just an empty
of the timetable and further it is explained in the flow slots which is further discussed in results section in
chart given in figure1.

115
table(i).Pseudo code of Algorithm for assigning teachers and Where x is number of day i.e. for Monday x=1 and for
courses are given below in which courses are divided into Tuesday x=2.
theory and practical:

TABLE I.
MONDAY
No of
Slots
Slots Subjects Teachers

1 08:00 to 09:00
EMPTY EMPTY

2 09:00 to 10:00
EMPTY EMPTY
3 10:00 to 11:00
EMPTY EMPTY
4 11:00 to 12:00
EMPTY EMPTY
5 12:00 to 01:00
EMPTY EMPTY
6 01:00 to 02:00
EMPTY EMPTY

Table i Results obtained from algorithm defined in pseudo code 1

The above table shows results of first algorithm in which empty

timing slots are created which is further filled by second
algorithm.
Similarly table ii given below shows the results obtained by
pseudo code 2 shown in figure 2, it shows how the algorithm
assigns the teachers to the slots created by 1st algorithm shown
in the pseudo code 1 defined in figure 1 whose results obtained
are shown in table i.

TABLE II.
No of MONDAY

Slots Slots Subjects Teachers

1 08:00 to 09:00 Computer Vision Ma’am

PR Shafiya
Figure 3 pseudo code 2 for assigning different teachers to empty 2 09:00 to 10:00 Computer Vision Ma’am
slots PR Shafiya
3 10:00 to 11:00 Computer Vision Ma’am
Results for above code will be generation of time-table with PR Shafiya
assignment of courses and teachers which is presented in 4 11:00 to 12:00 Software
result section in table (ii). Architecture Dr.Qasim
TH
IV. RESULTS 5 12:00 to 01:00 Web Technology Ma’am
TH Rabeea
To acquire the results of proposed model the system that has 6 01:00 to 02:00 Interactive Ma’am
been used in generation of timetable intel core i5 with 8gb Mulitimedia TH Memoona
Ram and 3.2 giga hertz processor the results of algorithm one
and two are given below in table i and ii respectively .These Table ii Results obtained from pseudo code 2 defined in figure 2
algorithms are designed by taking instances of Mehran uet
time-tables, this algorithm gives accurate results for the
constraints imposed in Mehran uet satisfying all the
requirements present in a time-tabling scheduling which is
comprised in a piece wise function

116
As the second algorithm runs allocation of all subjects is Annals of Operations Research), University of Nottingham,
done meeting the expectations of both hard and soft UK, School of Computer Science & IT, 2002.
constraints. [8] M. W. Carter, “A survey of practical applications of
The figure 4 given below acknowledges the parameters like examination timetabling algorithms,” Operations Research,
consistency , modifiability , reusability , error ratio and vol. 34, pp. 193–202, 1986.
accuracy which is calculated by [9]S. C. Brailsford, C. N. Potts, and B. M. Smith,
“Constraint satisfaction problems: Algorithms and
applications,” European Journal of Operational Research,
vol. 119, pp. 557–581, 1999.
[10]G. M. White and P. W. Chan, “Towards the
construction of optimal examination timetables,” INFOR 17,
pp. 219–229
[11]A. Colorni, M. Dorigo, and V. Maniezzo, “Genetic
algorithms - a new approach to the timetable problem,” in
Lecture Notes in Computer Science - NATO ASI Series,
vol. F 82, pp. 235– 239, Combinatorial Optimization,
(Akgul et al eds), SpringerVerlag, 1990.
[12]Salwani Abdullah and Hamza Turabieh," Generating
University Course Timetable Using Genetic Algorithms
and Local Search ",Third 2008 International Conference on
Figure 4 shows the achievement of algorithm used Convergence and Hybrid Information Technology.
[13]Suryanti Awang et al,"Generating University Timetable
V. CONCLUSION by using a Combination of Heuristic and Random Selection"
In this conducted research, an algorithm is purposed for ,
University course timetabling problem which is based on [14] Mei Rui ,Guan Jinyu and Li Bo,"University Course
Mehran Uet time-table and it's taken in consideration which Timetable System Design and Implementation Based on
satisfies all the constraints. Piece wise function is also Mathematical Model",
introduced which come up with the expectations of dividing [15] Anirudha Nanda, Manisha P. Pai, and Abhijeet
the slots according to Mehran Uet's timetable, another Gole,"An Algorithm to Automatically Generate Schedule
algorithm is introduced to allocate courses to those empty for School Lectures Using a Heuristic Approach "
slots. Our Future work depends upon reducing the error ratio. ,International Journal of Machine Learning and Computing,
Vol. 2, No. 4, August 2012.
[16] Anuja Deshmukh, Mayuri Kandalkar and Prof. Rashmi.
REFERENCES P. Bijwe,"Comparitive Study of Algorithms for Solving
[1] E. K. Bruke and S. Petrovic, “Recent Research Time Table Generation Problem",International Journal of
directions in Automated Timetabling”, European Journal of Advanced Research in Education & Technology
Operational Research, Vol.140, No.2, (2002), pp.266-280. (IJARET)Vol. 3, Issue 2 (April - June 2016)
[2] D. De Werra, “An introduction to Timetabling”, [17] D. De Werra, “An introduction to Timetabling”,
European Journal of Operations Research, Vol.19, (1985), European Journal of Operations Research, Vol.19, (1985),
pp.151-162. pp.151-162.
[3] R. Qu and E. K. Burke, “Adaptive Decomposition and
Construction for Examination Timetabling Problems”,
Multidisciplinary International Scheduling: Theory and
Applications, P. Baptiste, G. Kendall, A. Munier-Kordon, F.
Sourd, (Editors.), (2007), pp.418-425.
[4] Wren, Scheduling, “Timetabling and Rostering – A
Special Relationship!”, The Practice and Theory of
Automated Timetabling: Selected Papers from the 1st Int'l
Conf. on the Practice and Theory of Automated
Timetabling, E. K. Burke, P. Ross (Editors), Lecture Notes
in Computer Science, Springer Verlag, Vol.1153, (1996),
pp.46-75.
[5] Abdullah.S and Turabieh.H,"Generating University
Course Timetable Using Genetic Algorithms and Local
Search", Third 2008 International Conference on
Convergence and Hybrid Information Technology.
[6] A. Hertz, “Tabu search for large scale timetabling
problems,” European journal of operations research, vol. 54,
pp. 39–47, 1991.
[7]E. K. Burke and J. P. Newall, “A new adaptive heuristic
framework for examination timetabling problems,” in
Technical Report NOTTCSTR- 2001-5 (submitted to

117