IJCSNS International Journal of Computer Science and Network Security, VOL.10 No.

7, July 2010

237

Adaptive E-learning using Genetic Algorithms

Samia Azough, ,Mostafa Bellafkih and El Houssine Bouyakhf ,
National Institute of posts and Telecommunications Rabat, Morocco
Faculty of Sciences Rabat Agdal, Morocco
Summary
In this paper, we describe an adaptive system conceived in order
to generate pedagogical paths which are adapted to the learner
profile and to the current formation pedagogical objective. We
have studied the problem as an "Optimization Problem". Using
Genetic Algorithms, the system seeks an optimal path starting
from the learner profile to the pedagogic objective passing by
intermediate courses. To prepare the courses for adaptation, the
application creates a descriptive sheet for resources, in XML
format, while its integration in the database. Some experiments
are added to this paper.

Key words:
Adaptive e-learning, Genetic algorithms, LOM standard,
Resources Modelization.

1. Introduction
With the growth of the internet, E-learning has gained a
lot of importance in many fields. This can be explained by
the multitude of advantages that it offers. In particular, the
Web enables the user to access to various resources of
formation (image, audio, video, simulations, etc), with a
good quality and from various locations. E-learning
systems exploit these resources in order to organize
flexible formation schedules more adapted to the users'
activities and preoccupations.
Nevertheless, the majority of the existing formation
platforms are generally conceived as contents distribution
systems, with few care about the interests of singular
learners. Actually, E-learning concept suffers from many
inherent lacks:
the teacher is not permanently present;
the teacher is not directly confronted with the learner;
the management of the immediate reactions of the
learner is difficult.
The earlier efforts in this way were tutoring systems. They
were concerned with the interaction between the learner
and the system. They produced adaptive Intelligent
Systems (IS) able to generate didactic material according
to the total behavior of the learner. These systems allowed,
using Artificial Intelligent Techniques, creating a
completely interactive process adapted to the formation.
These systems are particularly conceived to meet the need
Manuscript received July 5, 2010
Manuscript revised July 20, 2010

of formalizing or of representing the course topics, the

learners and the pedagogic strategies. Among the main
suggested adaptations, we find educational system GAITS
[12]. In GAITS, supervised teaching techniques are
employed where the teacher assigns a pedagogical
objective to each student before teaching begins.
In order to achieve this objective, the tutor interacts with
the learner using predefined dialogues. Dialogues are
composed of a Header and a Body. The Header contains
prerequisite knowledge that must be mastered by the
student to consider this dialogue as a candidate, taught
knowledge that comprises the actual lesson, a learning
level, and a learning strategy to present the lesson. The
Body represents the teaching materials that define the type
of the interaction (e.g., exposition, question/answering,
game, etc.)[11].
Essential policies should be adopted in order to conceive
learner centric systems taking in consideration its own
needs. Asking the trainer to conceive course according to
each leaner profile and to each pedagogic aim and
demanding permanent managing of learner reactions are
possible solutions to achieve the adaptation, but they are
very expensive in terms of time and cost. Instead, the
following parameters should be improved to build an
adaptive system of e-formation:
profitability to produce more significant results in
terms of quality and costs;
reusability to reuse the already existing pedagogic
resources;
flexibility easy to implement at several various
platforms;
adaptability to adapt for several profiles of learners;
interactivity to manage the learner reactions during the
formation.
In our work, we have conceived an adaptive educational
system based on the modelization of the description of
pedagogic resources. This system will be able to provide
the path the most adapted to the learner profile by using
algorithms of optimization. For this reason, we propose to
model the description of pedagogic resources in a
descriptive sheet under XML format. This model will

238

IJCSNS International Journal of Computer Science and Network Security, VOL.10 No.7, July 2010

facilitate the search and the proposal for resources adapted

to the learner profile in progress and the pedagogical goal
of the formation. Providing an E-Learning platform with
adaptive utilities needs thorough reflection. The learner
must be guided when consulting his courses, however he
should be enabled to better understand his steps of
formation according to his capacities and thus to be able to
auto-evaluate. Hence, the adaptive e-learning will be able
to help the learner to be more autonomous, to have a better
comprehension of the course and also to better apprehend
and manage his learning process.
The article is structured as follows: In section 2, we
present the architecture used for designing the system to
be implemented. In section 3, we describe the
modelization module and we define multimedia Learning
Objects (LOs) as well as LOs metadata. The section 4
relates to the adaptation module, it defines Genetic
algorithms and adapter functionality. The section 5 present
the scenario implemented in modelization and adaptation
modules. Finally, we discuss results in section 6 and we
conclude in the last section.

2. System Architecture
2.1 Approach presentation
The current proposed adaptive systems require rebuilding
the courses according to a specific format defined by the
researcher. In instance, they use adaptive hypermedia [8].
In fact the trainer must construct new courses and he can
not reuse earlier developed courses. Additionally, the
system does not permit the trainer to choose the form
according to his experience. To satisfy such need, we
propose to model the description of the resources,
whatever are its formats, basing on the available
knowledge.
On one hand, the learner must own a precise intellectual
baggage to be able to understand a definite course. This
knowledge set, which is called pre-requisite concepts, is
commonly defined by either the author of the pedagogic
resource or an expert. On the other hand, each course is
associated to a definite knowledge set that should be
acquired after the course (pedagogic goal of the course);
we call them the post-acquired concepts (post-concepts).
Thus, the approach used to describe the course is based on
the study of the state envisaged of a virtual learner before
and after taking this course. To model the pedagogical
resources, we have built for each resource a describing
sheet based on the pre-requisite and acquired concepts
intervening in the latter.
From a semantic point of view, several concepts intervene
in the contents of the pedagogical resources, either as
preliminary knowledge (pre-requisite), or as knowledge to

be acquired (post-acquired). The pre-requisite ones and

post-acquired of each course will be presented in the sheet.
A good description of the contents makes possible to
present to the learner the course most adapted to his
profile.
To discern concepts intervening in each pedagogic
resource, we class them into modules. The resources
belonging to the same module deal with various
knowledge (concepts) according to the pedagogic
objective of the formation.
In order to define the module in the database, the expert
registers the various concepts approached by the resources
belonging to the same module of formation as well as the
relations between them. The system forms the concepts
tree and records it in the database as a characteristic of
each module of formation.
The choice of the description based on the concepts
enables us to remain independent of any pedagogic
approach and any resource format. The modelization of
the resources description [2] in XML format will allow
their reuse. It allows also the automatic design of various
pedagogic courses and thus the selection of the course
adapted to the learner profile [1]. The research of the
adapted courses is transformed into the optimization
problem of researching the most optimal path from a
starting point (learner profile) to a final point (pedagogic
objective) while passing by intermediate points (courses).
The result will be in a form of courses list which has to be
attained.

2.2 General Architecture

In this part, we present various diagrams of application
design for an adaptive e-learning. The objective is to
conceive a system which can model the description of
pedagogic resources and guide the learner in his formation
according to his assets and to the pedagogic objective that
is defined by the trainer. This pedagogic objective presents
the capacities that the learner must have acquired at the
end of the formation activity.

Fig. 1. Structure of the system

IJCSNS International Journal of Computer Science and Network Security, VOL.10 No.7, July 2010

Our application is made up of two principal modules.

Learner Space (1) (fig.1) accommodates the identifiers of
a learner, selects his profile from the Learners database
and returns it to the adapter as well as the goal of this
formation.
The adapter (adaptation process) uses optimization
algorithms to seek the optimal strategy while selecting the
courses in the resources base and provides them to the
user interface. The Learner database contains the
identifiers of the learner and his knowledge or asset. As a
result, the system provides an optimal courses list to
achieve the current goal by applying the genetic
algorithms to seek the intermediate states.
The part (2) related to the modelization of pedagogic
resources prepare them to be used by the adapter. In
Expert Space, the teacher or the expert, who seeks to
integrate new resources in the base, describes them by
filling a form.
Registered information will be recorded in XML sheet by
the modelisator (modelization process). The sheet,
recorded in the database, facilitates the reuse of these
resources. The pedagogic resources base contains the
courses that have to be attained. These courses are
associated with a descriptive sheet containing preliminary
knowledge called pre-requisite concepts and acquired
knowledge called post-acquired concepts.
This database is composed of modules; each module is
composed of several courses.

239

A function to be optimized which turns over a

function of adaptation (fitness) of the individual.

A mechanism of selection of the individuals allowing

statistically to identify the best individuals of a
population and to eliminate worse.

Operators allowing to diversify the population during

generations and to explore the space of the other
possible solutions. The operator of crossing (crossingover) recomposes genes of individuals existing in the
population. The purpose of the operator of change is to
guarantee the exploitation of the space of solutions.
Parameters of dimensioning such as cut population,
criteria of stop, probability of application of the genetic
operators.

3.2 Structure and Operation of the adapter

Sheets built during the modelization step and integrated in
the database are used to produce the path most adapted to
a learner profile. [1]

3. Adaptation
3.1 Genetic Algorithms: Definitions and Process
The genetic algorithms are algorithms of optimization
based on the theory of the evolution of the species of
Charles Darwin [4]. The first work of John Holland goes
back to the years 1960 and found a first result in 1975
with the publication of Adaptation in Natural and
Artificial Systems [6]. It is however the work of David
Goldberg who largely contributed to develop the genetic
algorithms [5]. The purpose of a genetic algorithm is to
seek an extreme of a function defined on a space of data.
The algorithm rests on the following points:
A coding principle of the population element; the
quality of the data coding conditions the success of the
optimum research. Two forms of coding are found: the
binary coding and real coding.

A mechanism of generation of the non-homogeneous

initial population. The choice of the population
conditions the speed of convergence towards the
optimum.

Fig. 2. Structure of the adapter

The adapter (fig.2) selects from the resources base the

courses which can be intermediate states by comparing
their sheets with the student profile and the pedagogic goal.
We transformed the search for adapted course into a
problem of optimization using the evolutionary algorithms.
The goal, the learner profile, the pre-requisite and postacquired concepts are seen at the entry of the adapter
under the format of vectors.
Goal (1 1 1 0 1 1) means that after following this
formation, the learner must have acquired the concepts (1,
2, 3, 5 and 6).
Profile (1 1 0 0 0 0) means that the learner has already
acquired concepts 1 and 2.
For the courses, the pre-concepts enable us to check the
conditions to reach the course in question. The postconcepts enable us to determine the probable state of the
profile of virtual learner after the follow-up of the course
in question.

240

IJCSNS International Journal of Computer Science and Network Security, VOL.10 No.7, July 2010

In the genetic algorithm, coding used is binary. The fitness

function is calculated according to adaptation of learner
for the courses under various formats. The probability of
crossing used is equal, and can be corrected by observing
the actions of learner and the evolution from its profile. It
means in our application the contribution of the capacity
and the intelligence of learner in acquisition of concepts.

3.3 Adaptation Process and Implementation

To integrate the genetic algorithms into the centre of our
system, we used package JGAP. JGAP is a Genetic
Algorithms and Genetic Programming Component
provided as a Java framework. It provides basic genetic
mechanisms that can be easily used to apply evolutionary
principles to problem solutions.
The idea is to transform the problem of searching for the
adapted path into an optimization problem. The starting
point is the learner profile, the arrival point is the teaching
goal of the formation and the intermediate states are the
evolutions of the profile after taking the available courses.
To implement the genetic algorithm we followed five
stages:
To predict our chromosome.
To implement a function of fitness.
To install an object of configuration.
To create a population of the potential solutions.
To evolve population!
We envisage evolving of individuals (courses) according
to the genetic operators to build a final individual solution
of our problem.
Chromosome Course: The chromosome course is
presented in the form of Boolean genes, its size is the
number of concepts on which the formation module is
based. The values of genes present the envisaged
profile of a virtual learner attaining the course in
progress.
Fitness Function: The function of fitness is
implemented according to the learner profile and to the
pedagogic goal of the formation. The individual
maximizes the function in question.
Genetic Operator: Starting from two individuals
courses, the operator produce an individual course
(solution) result of union of the two courses in
question. To follow the trace of the operator, we add a
gene to our chromosome in the form of a character
string (name of the generated individual) formed of the
concatenation of the names of the courses submitted to
the operator.

Object of configuration: We create a configuration

object with our fitness function, we initialize the
chromosome and we choose the size of the population
to evolve.
1) Population: The initial population is formed by
selecting from the database the courses which can
be used to reach the goal.
2) Evolution of the population To evolve the
population, we apply the designed genetic
operator and we post the solution which
maximizes the fitness function.

4. Modelization
4.1 LOM ( Learning Object Metadata)
The purpose of this multi-part Standard is to facilitate
search, evaluation, acquisition, and reuse of learning
objects, for instance by learners, instructors or automated
software processes. This multi-part Standard also
facilitates the sharing and exchange of learning objects, by
enabling the development of catalogs and inventories
while taking into account the diversity of cultural and
lingual contexts in which the learning objects and their
metadata are reused. LOM (Learning Object Metadata)
data element is a data element for which the name,
explanation, size, ordering, value space, and datatype are
defined in this Standard [7]. Data elements describe a
learning object and are grouped into categories. The
LOMv1.0 Base Schema consists of nine such categories:
1. The General category groups the general
information that describes the learning object as a
whole.
2.

The Lifecycle category groups the features

related to the history and current state of this
learning object and those who have affected this
learning object during its evolution.

3.

The
Meta-Metadata
category
groups
information about the metadata instance itself
(rather than the learning object that the metadata
instance describes).

4.

The Technical category groups the technical

requirements and technical characteristics of the
learning object.

5.

The Educational category groups the educational

and pedagogic characteristics of the learning
object.

IJCSNS International Journal of Computer Science and Network Security, VOL.10 No.7, July 2010

6.

The Rights category groups the intellectual

property rights and conditions of use for the
learning object.

7.

The Relation category groups features that define

the relationship between the learning object and
other related learning objects.

8.

The Annotation category provides comments on

the educational use of the learning object and
provides information on when and by whom the
comments were created.

9.

The Classification category describes this

learning object in relation to a particular
classification system.

4.2 Resources Classification by Modules

The current e-learning rests on a model of static modules
diffused via the Web. The course followed by learner is
fixed once for all by the teacher at the creation phase of
the module and, that, without caring of the preferences and
results obtained. To allow an automatic adaptation of the
module, we followed the process below (fig.3).

241

sheet based on the data provided by the user and records

the result in the base.
The content scenarisation makes sense of its structure and
determinates the arrangement of concepts approached by
the learner in his formation process. The resources format
or the pedagogy used while creating courses are not fixed.
The application let the teacher to choose the convenient
ones according to his experience. So that, he can add to
the database any pedagogic resources he thinks will be
useful to the student.
The integration of new resources in the database is done
by filling the form presented by the Integration interface
of new resources. The expert chooses the correspondent
module.
If the module to be chosen was already designed by an
expert college, a consultation of the module would be
necessary to remain in analogy with the definition of the
concepts presented and to be able to add others to them if
it is necessary. The scenarisation content is done following
the figure below (fig.4).

Fig. 4. Integration of the pedagogic resources

Fig. 3. Development process of the module

The cited products (fig.3) are implemented in the

application and the tasks are done by the expert (teacher).
To define a new module, concepts, which will intervene in
resources belonging to the module in progress, are
mentioned through a form.
Hence, the expert completes the form by the relations
between concepts. The stage of determination of the
concepts is carried out before the integration stage of new
resources in the base.
Out of the integration of the resources, the expert selects
among the concepts of the current module those
intervening in the resource in progress as pre-requisite or
post-acquired concepts. The system then produces the

When we determinate concepts intervening in the

pedagogic resource either as pre-requisite or assets, we
can easily fill the form. After the data acquisition, they are
provided to the system to build the sheet and to record it in
the database.

4.3 Modelization Process

In order to implement such system and to test it, we
choose a profile for applying the LOM standard. To
design the sheet, we use XML language and the LOM
standard which is most widespread in the description of
the resources.
The choice of XML to present the metadata indicates the
intention of opening and facilitating without predicting
technical choices of implementation to be used.

IJCSNS International Journal of Computer Science and Network Security, VOL.10 No.7, July 2010

242

The characteristics used to design the descriptive sheet of

the resource are:
identifier (LOM 1.1),
title, (LOM 1.2)
name(s) of the author(s),
dates of realization and last modification, (LOM 8.1 and
8.2)
key words, (LOM 1.5)
description (LOM 1.4)
format (LOM 4.1)
size of the resource, (LOM 4.2)
localization of resource (URL). (LOM 4.3)
language (LOM 1.3)
pedagogic type of resource (LOM 5.2)
pre-concepts (LOM 9.4)
post-concepts (LOM 9.1)
The sheet is designed according to the following format
(fig.5).

The sheet makes possible to describe several types of

resources and thus to compare and evaluate the various
formats and their impacts on the learner profile. To focus
on certain parameters allows to reduce the size of the data
at the treatment phase and consequently to minimize the
execution time of the application.
In the article [1], the sheet is considered as available for
each resource of the base. The aim was to be able to
propose the most adapted path for the learner profile. The
sheet is henceforth designed automatically during the
integration of new resources in the base. The modelization
part of our application makes possible to design the sheet
and to organize the database.
In this application two principal actors intervene:
The Learner selects a module. His knowledge is then
evaluated to update his profile. He will choose a
pedagogic objective among those predefined by the
expert. The research of the adapted path is made
thereafter and the result is in the form of a courses list
to follow. These courses are recorded in a file called
virtual portfolio to allow the consultation during each
connection to the application. During the formation,
the learner can test the evolution of his capacities
thanks to the tests attached to the attained courses.
The expert defines the modules of formation by filling
a form. He integrates the pedagogic resources into the
database. He attaches to each resource a kind of
marked exercise (tests) to allow the evaluation of
learner knowledge. Also, he defines pedagogic
objectives to guide the learner during his formation.

5. Experimentation Remarks
5.1 Profile Modelization
Fig. 5. An example of created sheets

The sheet is designed automatically starting from the data

provided by the expert. The following parameters mean:
num: identifier of the concept;
Name: name of the concept;
Value: relevance of the concept (level of boarding of
the concept: beginning/ intermediate /Advanced);
Type: pre-requisite or post-acquired concept.
The data recorded in the card can be used to index
resources and to facilitate the research and the automatic
construction of the relations between them.

Student level is considered in experimentation tests as

predefined data. To automatise the action of evaluating
knowledge students we propose to use the following
process. To get informed about the learner level, we
propose for a learner the tests to pass. These tests are in
fact added by the expert and attached to the resources to
allow the follow-up of the formation and the update of the
profile.

IJCSNS International Journal of Computer Science and Network Security, VOL.10 No.7, July 2010

243

By choosing an objective, the system calls the adapter for

using the genetic algorithms. The obtained result is a list
of courses to attain. These courses are copied in the
portfolio, and the linked exercises (tests) are put in the
auto-evaluate part to allow the learner to evaluate his
capacities.
The adapter return results such as presented in fig. 7. The
number of boolean genes in the chromosome is equal to
the number of current module based concepts.

Fig. 6. Stages followed for the modeling of the profile

The teacher proposes questionnaires allowing to evaluate

the learner knowledge and thus to direct his formation.
Out of the inscription, the learner selects a module,
answers the associated questionnaires and passes at the
stage of research of the adapted path.

5.2 Adaptation Result

Let us consider the sheet presented in fig.5. The block
which defines the concepts is used by the adapter for the
research of the adapted course. It is transformed in binary
format. For the course presented above (fig.5):
(1100000010000): pre-concepts or preliminary state of
the profile of a virtual learner who will be able to
follow this course.
(1100000010001): post-concepts or preliminary state
of the profile of a virtual learner who has followed this
course.
These two vectors will be supplemented by using the tree
of concepts of the module in question.
The use of such sheet makes possible to consider its
integration in several interfaces and the extension towards
other characteristics (or attributes) of the resources. It can
also be used to describe the contents of different form
resources either than the courses (exercises, examinations,
articles bound, software,) by adding other attributes. The
design of the sheet in format XML makes possible to
consider a regeneration of a meta-sheet with other
attributes making possible to study quality of pedagogic
resources. The attribute Value, used to indicate the
relevance of the concept in the course, can be equal to a
percentage by adding to the application a fuzzificator
(fuzzy logic) to define the rules.
Through the Student Interface, the learner can choose an
objective for seeking adapted path. He will consult his
virtual portfolio and write messages. In order to evaluate
his assets, he can access to tests attached to resources
contained in his portfolio.

Fittest Chromosome is: Size:19, Fitness value:50.0,

Alleles:[BooleanGene=true,
BooleanGene=true,
BooleanGene=false, BooleanGene=true,
BooleanGene=true, BooleanGene=true,
BooleanGene=false, BooleanGene=true,
BooleanGene=true, BooleanGene=false,
BooleanGene=true, BooleanGene=false,
BooleanGene=false, BooleanGene=false,
BooleanGene=false, BooleanGene=false,
BooleanGene=false, BooleanGene=false,
StringGene=C6C10C16]
Fig. 7. Solution obtained using the adapter

The StringGene (fig.7) indicates the path followed to

obtain the fittest chromosome. This string is transformed
to a list of courses to take by using indexes.
Until now, the results dont care of the student feedback.
Actually, we try to perform results through fitness
function. We are looking for a function which can choose
courses according to last feedback of the learner in
progress.

5.3 Application
In order to examinate the results of this approach, the
expert defined a module through the application interface
following the steps below.
-

Definition of based concepts of the module in

process;

Definition of courses;

Defintion of pedagogical objectives of the

formation.

All this data are introducing to the database through a

form. After defining the module of formation, the learner
can begin the adaptation process by choosing an objective
from those defined in the database. Then, the adapter will
use the genetic algorithm and generate an adaptive path to
follow to attain the chosen objective

244

IJCSNS International Journal of Computer Science and Network Security, VOL.10 No.7, July 2010

6. Conclusion
The application was conceived to assistance of free tools
in the form of dynamic Web site. That facilitates its use
and reduces the finances. The use of the genetic algorithms
enabled us to automate the search for courses adapted to
the learner.
Thus, we do not solidify the pedagogic course to take.
The fact of modeling the description of the resources
makes possible to benefit from the pedagogic experiment
of the expert (teacher) in designing the courses. Separation
between the contents and its presentation allows the
application of generic models and the integration of
specific pedagogic models, also the use of XML format
makes possible to really consider the generalization of the
logical structuring of numerical information.
In prospect, we will define and implement the rules of
fuzzy logic to treat the levels of a beginner learner,
intermediate or expert in the acquisition of the concepts.
We also hope to study the user feedback and the impact on
the profile of the learner.

References
[1] Azough S., Bellafkih M. (2007) Reuse of resources for
adapted formation to the learner. 3rd International
Symposium on Computational Intelligence and Intelligent
Informatics ISCIII IEEE Agadir, Morocco, March 28-30.
[2] Azough S., Bellafkih M., Bouyakhf El H. (2008) Elearning adaptatif : modlisation des ressources en XML,
CINQUIEME CONFERENCE SUR LES SYSTEMES
INTELLIGENTS : THEORIES ET APPLICATIONS
(SITA08),
Institut
national
de
Postes
et
Tlcommunications Rabat, Maroc, 5-6 Mai.
[3] Azough S., Bellafkih M., Bouyakhf El H. (2009) Elearning adaptatif :
modlisation des ressources et
adaptation au profil de lapprenant, Systmes Intelligents:
Thories et Applications Edits par europia Productions.
ISBN:978-2-909285-55-3. pp51-68.
[4] Darwin Charles (1980) Lorigine des esp`eces, petite
collection maspero, Paris.
[5] Goldberg, David Edward (1989) Genetic algorithms in
search, optimization, and machine learning, AddisonWesley, 412 p.
[6] Holland J.H. (1992) Adaptation in natural and artificial
systems: an introductory analysis with applications to
biology, control, and artificial intelligence, Cambridge,
Mass. : MIT Press.
[7] LTSC-IEEE LOM Working Group (2001) Draft for
Learning
Object
Metadata
(LOM
6.1),
http://ltsc.ieee.org/wg12/
[8] Michel C., Rouissi S., (2002) Etude de l'organisation et
caractrisation de l'information pdagogique pour construire
des documents hypermdias adaptatifs diffuss sur le web,
5me Colloque International sur le Document lectronique
(CIDE05).Hammamet, Tunisie. 308p - pp153-167.
[9] Quafafou, M. GAITS: Fuzzy Sets-Based Algorithms for
Computing Strategies Using Genetic Algorithms.

Proceedings on Fuzzy Logic in Artificial Intelligence FLAI

1993, in The Lecture Notes in Artificial Intelligence,
Springer Verlag.
[10] Semet, Y. Lutton, E. Collet, P. Ant colony optimization for
E-learning: observing the emergence of pedagogic
suggestions. Swarm Intelligence Symposium, SIS 03.
Proceedings of the 2003 IEEE. 24-26 April 2003.

Samia Azough born in Casablanca, Morocco. She received the

Research Master's in Computer Engineering, Telecoms and
Multimedia from Mohammed V University of Rabat in 2003.
She is conducting, since 2004, doctoral research at e-learning
systems, modelization of pedagogic resources and adapting
formation to the learner profile.

Mostafa Bellafkih born in Oujda, Morocco. He received the

PhD thesis in Computer Science from the University of Paris 6,
France, in June 1994 and Doctorat Es Science in Computer
Science (option networks) from the University of Mohammed V
in Rabat, Morocco, in May 2001. His research interests include
the network management, knowledge management, A.I., Data
mining and Database. He is Professor in The National Institute of
Posts and Telecommunications (INPT) in Rabat, Morocco since
1995.