See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/356075614

Competenty-Based Approach for Learning Objects sequencing using DNA

computing

Conference Paper · June 2021

DOI: 10.1109/ICDATA52997.2021.00030

CITATIONS READS
0 10

2 authors:

Amina Debbah Yamina Mohamed Ben Ali

Badji Mokhtar - Annaba University Badji Mokhtar - Annaba University
3 PUBLICATIONS   6 CITATIONS    48 PUBLICATIONS   125 CITATIONS   

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Amina Debbah on 22 August 2022.

The user has requested enhancement of the downloaded file.

 2021 International Conference on Digital Age & Technological Advances for Sustainable Development (ICDATA)

Competenty-Based Approach for Learning Objects sequencing using DNA computing

2021 International Conference on Digital Age & Technological Advances for Sustainable Development (ICDATA) | 978-1-6654-2901-6/21/$31.00 ©2021 IEEE | DOI: 10.1109/ICDATA52997.2021.00030

Debbah Amina Mohamed Ben Ali Yamina

Computer Science departement Computer Science departement
Annaba, Algeria Annaba, Algeria
[email protected] [email protected]

Abstract—In the e- learning systems, a learning path is route between two nodes. As a result the CS problem falls in
known as a sequence of learning materials linked to each the NP complete class of problems [4]. DNA computing has
others to help learners achieving their learning goals .As it is shown its efficiency for solving NP hard problems such as
difficult to have the same learning path that suits different spanning tree, Travel Salesman Problem (TSP), Hamiltonian
learners, the Curriculum Sequencing problem (CS) consists of path Problem (HPP). In our study we propose to use the
the generation of a personalized learning path for each learner DNA computing method for solving the sequencing
according to his learner profile. This last one is represented problem (optimal route/path) through competencies in the
through competencies. The proposed approach includes two e-learning system for pedagogical sequencing. It allows
components: (1) competency based approach that is used to finding the optimal path for each learner profile in order to
represent the knowledge model and to deliver to the learner adapt the learning materials through competencies . The
the learning path; and (2) a DNA computing approach based
paper is organized as follows: Following this introduction,
on a weighted graph is used to deliver a learning path to each
learner. An example regarding the course of algorithmic is
Section 2 presents the competency based learning objects
presented to demonstrate the effectiveness of the proposed sequencing by defining both learning objects and
design. Results show that competency based DNA computing competencies and their different specifications. Section 3
approach can generate a personalized learning path describes the DNA computing algorithm. And Section 4
successfully. presents the proposed Competency based DNA computing
Keywords- Learning path; Learner model; Domain model; approach for learning objects sequencing Finally, we present
Competency; Curriculum Sequencing Problem; DNA a full exemple from a real course use case to demonstrate
Computing; and validate our prposal, and discuss our conclusions for
this work.
I. INTRODUCTION II. COMPETENCY-BASED LEARNING OBJECTS
The adaptive e-learning systems contain a lot of learning SEQUENCING
materials that are digitalized and interchanged; these The use of the learning objects is widely employed since
learning materials are called “learning objects”. A learning it ensures a high level of interoperability and flexibility. The
object is defined as “any digital resource that can be reused systems based on the share and the reuse of learning objects
to support learning” [1]. The e-learning systems are are called LMS (Learning Management System) which their
characterized by self-guided learning and limited roles is to deliver the learning objects to the learners who are
instructor/trainer support [2] . Most e-learning systems connected to the system. The sequencing problem consists
provide a one size fits all environments where all the learners of constructing a suitable sequence of competencies and
are treated the same way in terms of learning materials. The learning object in order to be later delivered to the
personalization is an important feature of current century learners.These learning objects and competencies are then
and it is considered as the center stage for an effective assembled and/or aggregated in order to create greater units
learning experience. It consists of adapting the knowledge of instruction (lessons, courses, etc) [1]. Indeed the task of
according to learner’s needs such as preferences, sequencing was performed by the teachers who do not create
competencies and requirements. These needs are embedded a personalized sequence but instead a generic courses or
into the learner model. For the purpose of the lessons targeting generic learner profiles [5].The presented
personalization comes the idea of delivering to each learner a conceptual data model in this section describes for learning
personalized learning path known as the curriculum objects sequencing throughout competencies.
sequence. The Curriculum Sequencing problem (CS)
consists “to generate an individualized course for each A. Learning objects
student by dynamically selecting the most optimal teaching A learning object is defined, by IEEE(Institute of
operation (presentation, example, question or problem) at Electrical and Electronics engineers), as ‘‘an entity, digital
any given moment” [3]. In fact, the structure of the domain or non-digital, that may be used for learning, education or
model consists of graph. training’’. In order to allow the re-use of learning objects,
The point here is there is no known algorithm which they have to be annotated with metadata (e.g. author,
works in polynomial time for identifying a graph’s optimal keywords, description, etc). Metadata is supposed to enable

978-1-6654-2901-6/21/$31.00 ©2021 IEEE 109

DOI 10.1109/ICDATA52997.2021.00030

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on August 22,2022 at 12:16:30 UTC from IEEE Xplore. Restrictions apply.
reuse of these chunks by detailing the conditions of their instance of a metadata schema SCORM and that is stored
initial deployment [6]. Indeed many metadata schemas are and indexed into a Learning Object Repository. The learning
being used to describe these learning objects. We describe objects contains metadata that explain these learning
each of these standards by chronological order as follows: objects representing learning concepts (courses,lessons) in a
AICC stands for the Aviation Industry CBT domain model. We present our learning objects as delivered
[Computer-Based Training] Committee (AICC), which is elements to learners in the system. The figure below shows
an international association of technology-based training a description of the relation between the RCD and two
professionals which is able to support the annotation of respective learning objects in our approach.
learning objects.It was created in 1988. Since IMS focuses
on the interoperability of applications and online services
using metadata in 1999 the Sharable Content Object
Reference Model (SCORM) which is a model, not a
standard have been created. It attempts to allow reuse of
resources. Finally the Learning Object Metadata (LOM)
standard was approved by the IEEE (IEEE LOM, 2002).

B. Competencies
The competency is defined by [7]
as:”multidimensional, comprised of knowledge, skills and Figure 1: Competency attached to learning objects .
psychological factors that are brought together in complex
behavioral responses to environmental cues”. And as ”the Actually, we used the RDCEO specification in order to
integrated application of knowledge, skills, values, model competencies and According to RDCEO and IEEE
experience, contacts, external knowledge resources and nomenclature, a competency record is called ‘Reusable
tools to solve a problem, to perform an activity, or to handle Competency Definition’ (RCD).
a situation” . According to [7] the Competencies need to be
formally modeled so as to be able to interchange RCD Competencies are linked to the learning objects
competencies description between systems. For this via metadata records as a prerequisite or a learning
purpose, international specifications for competency outcome. The same competency could be a learning
description, such as the IMS RDCEO (Reusable Definition outcome for a learning object and a prerequisite for another
of Competency or Educational Objective) (IMS RDCEO, one, so a constraint is made between these learning objects
2002), the IEEE RCD (Reusable Competency Definitions) is the first one should be learned before the last one [10].
(IEEE RCD, 2004) and the HR-XML Competencies
(Measurable Characteristics) (HR-XML, 2006) are being A. The proposed model
used to model them. Among other features, all the standards In this sub section, the different parts composing the
offer records in order to model the competencies. proposed e-Learning system and their relationships are
presented. The most important part of the e-Learning system
III. DNA COMPUTING is the path generator component which allows to
automatically generate a sequence of Learning (i.e. a course,
DNA computing for (Deoxy Ribonucleic Acid) is a module or a lesson structured as a sequence of Learning
referred to as bio-molecular computing or molecular paths represented by Learning Objects and competencies)
computing, is also a new computational paradigm that during the learning process according to the learners profile
(personalization process). From a conceptual viewpoint the
employs molecule manipulation to solve hard
two sets of repositories are the learning objects repository,
computational problem instead of the traditional silicon
which consists of all the machine-understandable
based computer technologies. The Research in this field has representations of the relevant learning objects and the
been invented by [8] who made the first bio-molecular competencies repository containing all the competencies in
computation for solving difficult computational problems relationship with the learning objects as described above.
such as the Hamiltonian path problem using methods of And the last component is the path generator component in
molecular biology [4] , [9] and manipulating DNA strands. which the algorithm of DNA computing acts in order to
DNA is composed of two chains of repeating nucleotides. execute the personalized e-learning building process.
Each nucleotide consists of three components: a Phosphate Teachers interact with the learning objects repository and the
Group and two deoxyribose sugar and nitrogen containing competencies repository, providing artefacts (e.g Learning
four bases. objects, competencies) Learners interact with the
competencies repository in order to select one or a set of
IV. COMPETENCY-BASED DNA COMPUTING FOR objective competencies. So as to achieve the expected
LEARNING OBJECTS SEQUENCING adaptation capability, the path generator component uses
In our approach, a learning object is a learning content two specific models: the domain Model and the Learner
delivered through a Web Browser, that is annotated with an profile used to define personalized e-Learning sequences.

110

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on August 22,2022 at 12:16:30 UTC from IEEE Xplore. Restrictions apply.
The two sub models taken into account are, respectively: the repository that consists of courses, contents, pages or nodes,
knowledge about educational domain, the learner’s skills, and navigation links related to the design structure of the
and finally, the learning goals. In a personalized e-Learning represented data. The most important aspect of the domain
experience, these modalities depend, above all, upon these model is the relations between the course elements in other
selected educational domains, the learner’s learning skills. In word the navigation links between pages which are used to
the following subsections, we give detailed descriptions of decide upon adaptations. In our case the domain model
all the componenents of our proposed model. consists of learning objects which represents nodes. The
aggregation of these learning objects constitutes lessons and
courses. The relations between the learning objects are
represented by competencies as prerequisites and learning
outcomes. In order to construct the graph representing the
domain model, we need to explain how the learning objects
are related to competencies within the system. Indeed
learning objects metadata contains competencies
information such as the purposes of the competency namely
“prerequisite” or “learning outcome” that are the same used
in the competency record RDC. For instance we define a
learning object where 0 ≤ j ≤ n and a
competency where 0 ≤ i ≤ x . A competency weight
W is a value in the range {0,1} between the learning object
and the competency, 1 testifies that the learner’s competency
mastery rate is well mastered for the learning object,
while 0 declares the absence of the knowledge about the
competency in relation with the learning object. A sequence
is composed of a set of learning objects and competencies,
Figure 2: The architecture of the proposed model .
and eventually a set of arcs. Indeed each arc between two
learning objects and competencies has a weight which
B. Learner profile represents mastery rate of the competency as a learning
The learner is the main actor of the whole cognitive outcome or a prerequisite [12]. We formalize the above
process and he/she has to be represented in a described features as follows:
machine-understandable way in order to support the
personalization process of the e-Learning experience [4]. W(LO , Cm ) = ( ) (1)
The learner profile contains information about the learner. It Where Lo is a learning object and Cm is a competency
represents the learner’s properties and the way we represent and W is the weight.
them [11]. In our approach the learner profile consists of a The learning optimal path proposed to each learner
set of skills represented as competencies and other learner includes asp ecific set of learning objects and
properties such as e-mail address, the affiliation (student, competencies.To solve this problem, the system has to find
high school student), The measurable value of the the optimal sequence by maximizing the total weights (W)
competency represents the mastery rate where the value 0 represented by equation 1 on the weights path in order to
indicates a complete absence of acquired competency, while deliver to the learner the optimal path.
,
1 indicates a complete knowledge about the competency. (2)
The assessment process is realized by several kinds of tests max f(W) = W(LO , Cm )
(e.g. multiple choice test, etc.) that are embedded into the
system and used to assess learner knowledge about studied
competencies. Indeed a pre-assessment process is realized in
order to store the preliminary learner’s knowledge about the
domain model. After that the learner profile is updated We can formulate the sequencing problem as a
through a post assessment process. The utilization of the combinatorial optimization problem where for a given graph
competency is encouraged due to its usefulness for of learning objects and competencies, the purpose is to find
modeling the prerequisites and learning outcomes for the the optimal route leading to the learning goal between a start
learning objects, it is also for modeling learner current node and an end node by calculating the maximum cost path.
knolwedge .
C. Domain model V. RESULTS AND DISCUSSION
The domain model is one of the most important Results obtained using the DNA computing algorithm to
components used by adaptive e-learning systems to generate solve the Competency-based Curriculum sequencing
personalized courses for the learners. It represents a specific problem, we test it in a use case where the weighted graph is
domain of interest. The domain model serve as a data used to represent the domain model.

111

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on August 22,2022 at 12:16:30 UTC from IEEE Xplore. Restrictions apply.
 TABLE I . ACRONYMS OF LEARNING OBJECTS. In addition the table 5 represents the DNA codes of the
weights.
Learning object Acronym
TABLE IV. DNA CODES OF WEIGHTS .
Data Lo1
Statement Lo2 weights DNAcodes
Parameters Lo3 W1=0 CATG
Algorithm Lo4 W2=0.1 ATACGGAC
Program Lo5 W3=0.4 CGTAGAGTAACG
Operators Lo6 W4=0.5 AATTCTGACCCCAAAA
Expressions Lo7 W5=1.0 ACTATTTCTTACGTGGGCGT

The learner profile of the learner is represented mainly Step 2

by the competencies acquired by the learner and respectively
their mastery rates. The Table 2 below shows the details. Following the DNA encoding nodes step, we form the
Indeed, the construction of the weighted graph is done DNA strands of arcs between two nodes, we join the half
based on the evaluation of learner’s knowledge about each right of the first learning object with the corresponding
competency. half left of the second learning object to form the DNA
code of the weight between the two learning objects.
TABLE II .REPRESENTATION OF THE LEARNER
PROFILE
Step 3
Competency name Mastery rate Objective
competency Following the DNA encoding nodes step, the ligase
Variable and 0.4 Call a procedure operation is performed in order to get all the paths in the
constant sub-graph.
State a function 0.1
Local and global 0.1 Step 4
parameter
State a procedure 0.5
After amplifying the paths that begin with the right
Function and 1
proceudre starting node and the right ending node in the sub graph,
we obtain the results below.
In order to obtain the optimal sequence leading to the
learning goal, it is crucial to execute the steps of the DNA Step 5
computing algorithm.
The result of the final step corresponds to the solution to
Step 1 the Competency-based Curriculum Sequencing
After execution of the first step of the DNA computing Problem. The first proposed lesson is: variable and
algorithm, the encoding nodes results are generated constant-->Parameters--> local and global
randomly. The Table 3 below represents the DNA codes parameters-->algorithm-->state a procedure--
of the learning objects and competencies. >program-->call a procedure , and the DNA strand is :
CACGTAGAGTAACGACGAATACGGACGAC
GATACGGACATCCAATTCTGACCCCAAAAGAGA
TABLE III .DNA CODES OF LEARNING OBJECTS AND AATTCTGAC CCCAAAACGGCCATGGG
COMPETENCIES The second proposed lesson is: variable and constant--
>algorithm-->state a function-- >program-->call a
Learning objects / DNA codes DNA procedure. The DNA strand is:
competencies Complements
State a function TCTT AGAA CACGTAGAGTAACGATCCATACGGACTCTTATACG
Local and global GACG CTGC G ACCGGCCATGGG
parameters
Variables and AACA TTGT The first lesson represents the biggest DNA strand,
constants where the learner has more knowledge about the
Algorithm ATCC TAGG
program CGGC GCCG
competencies in this path, so we suggest this path to him
parameters ACGA TGCT because it is less consuming in terms of learning time.
Call a function TGGG ACCC Therefore the second solution is represented by the
Call a procedure GGAG CCTC smallest DNA strand by maximising the formula (2). The
Function and AGAT TCTA learner in this case has minor knowledge about the
procedure
competencies, so we suggest this path to him/her in order
State a procedure GAGA CTCT
to improve his mastery rate.

112

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on August 22,2022 at 12:16:30 UTC from IEEE Xplore. Restrictions apply.
 VI. CONCLUSION [5] L.De-Marcos, C. Pages, J.Martinez José and A. Gutierrez J,
“Competency-based learning object sequencing using particle
Creating a Curriculum sequence that conforms to each swarms”. In: Proceedings of the 19th IEEE International
learner in an e learning system is considered as a real Conference on Tools with Artificial Intelligence. pp. 111–116, 2007.
challenge. It plays a central role in the adaptive e learning
systems since the success of such systems depends on their [6] W.Chen,, Y. Hayashi, L.Jin, I. Mitsuru ,and R. Mizoguchi, “An
ontology-based intelligent authoring tool.” In: Proceedings of the 6th
ability to suit the needs of the learners. This study utilizes a international conference on computers in education. Beijing, China,
competency-based DNA algorithm approach to create the pp 41– 49. 1998.
Curriculum Sequences suited to each learner by calculating
the maximum cost path in the graph. It will enable the [7] G. Demetrios, Sampson, “Competence-related Metadata for Educational
learner in the first alternative to browse learning path of Resources that SupportLifelong Competence Development Programmes ,
Competence-related Metadata for Educational Resources that Support
learning objects in order to discover new competencies in Lifelong Competence Development Programmes”. Educational
the path. Therefore in the second alternative it will facilitate Technology Society, Vol. 12 (4), pp 149–159 2009.
the route until the learner reaches his learning goal. Besides
all the techniques used in the DNA computation the main [8] M.Leonard ,Adleman, “The manipulation of DNA to solve
one is the encoding method because it determines the mathematical problems is redefining what is meant by
optimal solution. In our own encoding method we have “computation”.,Scientific American Magazine 1998.
based essentially on the weight’s code to determine the
longest and the shortest strand. The DNA algorithm for [9] P. Brusilovsky , P.Miller, “Course Delivery Systems for the Virtual
solving the curriculum sequencing problem described above University.” In T. Tschang, & T. Della Senta (Eds.) Access to
was developed and implemented using the object oriented Knowledge: New Information Technologies and the Emergence ofthe
paradigm and java as an environment. Virtual University, , pp 167-206 2001.

[10] L. De-Marcos ,J. Martinez J, and JA. Gutierrez,” Swarm intelligence

in e-learning: a learning object sequencing agent based on
REFERENCES competencies”. In: Proceedings of the 10th Conference on Genetic
and Evolutionary Computation ,pp. 17–24 , 2008.
[1] D. Wiley, “ Connecting learning objects to instructional design
theory: a definition, a metaphor and a taxonomy” ,In: Wiley DA [11] K. Seki,T. Matsui, and T.Okamoto, “An adaptive sequencing method
(ed) The instructional use of learning objects. of the learning objects for the e-learning environment”. Electronic &
http://reusability.org/read/ chapters/wiley.doc.Accessed 2010. Communication in Japan, Vol .88(3),pp 54–71, 2005.

[12] A. Debbah, Y.Mohamed Ben Ali, “solving the curriculum sequencing

[2] C. Tattersall, J.Manderveld, B. den Berg, R. van Es, J. Janssen, and problem using DNA computing approach”, vol. 12(4), pp 1-18, 2014
R. Koper, “Self organizing way finding support for lifelong
learners.Educ Inform Technol” vol. 10, pp 109–121 ,2005.

[3] M.Huang, H. Huang,M. Chen,”Constructing a personalized

e-Learning system based on genetic algorithm and case-based
reasoning approach”. Expert Syst Appl vol. 33,pp 551–564,2007.

[4] G.Acampora, M. Gaeta, and V.Loia, “Hierarchical optimization of

personalized experiences for e-Learning systems through
evolutionary models”, Neural Computing & Applications, vol 20 (5),
pp 641-657,2009.

113

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on August 22,2022 at 12:16:30 UTC from IEEE Xplore. Restrictions apply.
View publication stats