Toward a Domain Ontology for Computer Projects Resolution:

Project Memory Challenge

Raja Hanafi

, Lassad Mejri

and Henda Hajjami Ben Ghezala

National School of Computer Science, University of Manouba, Manouba, Tunisia

Faculty of Science of Bizerte Carthage University, Bizerte, Tunisia

National School of Computer Science University of Manouba, Manouba, Tunisia

Keywords: Project, Computer Project, Project Memory, Knowledge Capitalizing, Ontology, Domain Ontology.

Abstract: In recent years, the project management has been practiced in many special computer projects which took

place in large companies. During the resolution of a project especially during the design phase, the project

leaders have encountered many problems which are treated and solved in the already existing projects. The

resolution of a similar new project forces project leaders to spend a lot of time accessing and reusing

existing project knowledge. This is why the problem of capitalization of knowledge proves to be very

important in order to solve the problem of time, of cost and of quality that a project manager can encounter

during his resolution. The best solution is to propose a technique for memorizing and saving knowledge.

This solution presents in a way the project memory. In literature, there are several approaches that are all

about the capitalization of knowledge and the construction of project memory. All these approaches are

generic models which are applied to any type of project such as the industrial and the technical project. In

this paper, we present a model approach for a project memory. In practice, this challenge is addressed by

proposing the domain ontology that characterizes the specification of computer project.

1 INTRODUCTION

According to the remarkable evolution of

technological life, sharing information and

experience between the actors of each organization

has been developed rapidly. Indeed, designers have

encountered problems during the design of their

projects. So, they don’t only use the shared

information to solve problems but also to avoid past

mistakes. Then, the proposition of a solution to

memorize tasks, actions and results during a finished

design project is proved to be fundamental.

In this context, we propose an approach for the

capitalization of computer project memory

knowledge.

This approach presents a decision support in the

project management phase of the design on the

previous plan, and this by proposing solutions and

problems that are already encountered in the

previous projects.

Our decision-making process will not only help

structure formalize and capitalize knowledge about

the resolution of a past project, but above all provide

a dashboard which is in the form of indicators,

information and a guide favoring the decision

making by leader of computer design project.

In the case-based literature (Benoît Eynard,

2001)(Paula Potes Ruiz, 2012)several architectures

are presented. Inspired by this work, we define the

architecture of our system on two main processes: an

off-line process and an on-line process.

In this paper, we focus on the first process

"offline process". This process concerns the

formalization of knowledge and the implementation

of domain ontology through three models: project

class model, project model and rational design

model.

Indeed, ontology is defined in computer science,

and the field of knowledge engineering (IC), as a

particular artefact to represent knowledge. It is now

classically accepted to distinguish the three levels of

ontology which are: (Hernandez N. M., 2007)

• The top-ontology, the highest level structuring

and the high-level knowledge '.

• The core or core ontology, provides the

concepts structure of the domain and describes

the relations between these concepts.

• The ontology of the domain, that is to say, the

concepts of the field as they are handled by

professionals.

Hanaﬁ, R., Mejri, L. and Ghezala, H.

Toward a Domain Ontology for Computer Projects Resolution: Project Memory Challenge.

DOI: 10.5220/0006936502470254

In Proceedings of the 10th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2018) - Volume 2: KEOD, pages 247-254

ISBN: 978-989-758-330-8

247

The paper is organized into three main sections.

Section 1 presents "related work" the main works

existing in the literature as well as a comparative

study. In Section 2, we reveal our proposed

approach by focusing on the offline process. The

third and the last Section introduces the notion of

ontology and the construction of the domain

ontology that we proposed.

2 RELATED WORKS

Mining the state of the art, various works addressed

the project memory models which aim at the

capitalization of knowledge and the construction of

project memory.

Zacklad and al (Zacklad, 2014), propose a

groupware "MEMO-net" using the DIPA problem

solving method for capitalization and knowledge

management in design projects. This groupware

makes it possible to manage the knowledge used in

order to better capitalize and reuse them. This

groupware is a tool that has two modules: (design

and diagnostic) that allows a project group to solve

problems encountered during the design

(capitalization of the design logic) and to preserve

the characteristics related to such a product. Ermine

(Ermine, 2001) has described its knowledge

management processes through the margerite model.

These processes can be internal or external. What

interests us is the internal process of capitalization

and sharing of knowledge within the company.

Serrano in (Serrano, 2014), proposed a global

system of capitalization of knowledge allowing the

actors of the company to exploit the important mass

of information. This system also makes it possible to

capitalize events in the field of OSI (Open Source

Intelligence) based on the Web Lab platform.

Other approach is proposed in (Harani, 1997) as

a design assistance tool whose main objective is the

capitalization of knowledge involved in the design

of a product for reuse.

Bekhti(Bekhti, 2003) proposed a dynamic project

definition and reuse process "DYPKM". This

approach is based on a method that provides a

structured trace of a project memory containing the

context in which the design takes place and the logic

resolution.

2.1 Comparative Study of the Studied

Approaches

Several classifications of project memory models are

available in literature. Inspired from these, we

present our own classification in the following table

(table 1) .This comparative study is based on a set of

criteria, namely:

• Simplicity of the Method: This is a primary

criterion because any method, as interesting as it

is, loses much of its value if its use is

complicated. Actors, who apply a knowledge

capitalization method, during the realization of

a design project, must not be obliged to acquire

new specific skills to be able to use this method

• Resource: To represent all the knowledge

forming the context of a design project, we need

as resources all the project management data. It

corresponds to the workspace, the data

representing the constraints to be considered

and the data of the project organization.

• Application Domain: This criterion gives a

global vision on the field of application of each

knowledge capitalization process or approach.

we have to work on this criterion because we

will focus, in our study, on computer projects.

Table 1: comparative study of some knowledge Capitalization models.

Model

Simplicity of the

method

Resource

Application

domain

Using of

(CBR)

Capitalization

level

Ermine's model

(Ermine, 2001)

Complex

(marguerite model)

Corporate memory Area of economy No Design

Zacklad ‘s

mode(Zacklad,

2014)

Complex

Collective

Software (DIPA)

Diverse

For all design

projects

Conception

and context

(Serrano, 2014)

Global + wave

(weblab platform)

Open source (blog,

internet, site ...)

Field of defence No Event

Harrani Model

(Harani, 1997)

Simple help tool

Company

knowledge

Computer,

mechanical,

industrial

No Design + Feature

Bekhti model

(Bekhti, 2003)

Simple process Project memory

Design project

(all areas)

Context + design

Rational

KEOD 2018 - 10th International Conference on Knowledge Engineering and Ontology Development

248

• Using of CBR (Case Based Reasoning): we

have introduced this criterion to check whether

the concept of case-based reasoning is used in

the proposed model or not.

• Capitalization Level: this criterion allows us to

define the part that a process or approach can

capitalize.

2.2 Discussion

We have presented in this comparative study a set of

models that help to capitalize knowledge. It is a

clear and detailed study that allows us to note that:

• All the models proposed are almost all complex

models which are based on other methods. The

user of these models must have an additional skill

to be able to use these processes of capitalization.

• For these models and others existing in the

literature, the consideration of project memory as

a knowledge resource is almost totally absent.

• Almost all the proposed models are applied for

design projects and this has shown us the

importance of these types of projects.

• Finally, we have noticed the absence of a model

which guarantees the capitalization of all these

concepts at the same time which are: project

context, project characteristics and rational

design.

• Based on this comparative study we will propose

in the next section our approach to capitalize

knowledge of project memory. This approach

aims to provide decision support in project

management from the design phase to the

previous plan.

3 PROPOSED APPROACH

Our goal is to present an approach to help the leader

to solve its new project by referring to the

experiences and knowledge which are stored in a

project memory. This section, introduces the

architecture of our approach and in particular the

modeling part which composed of three models: the

project class model, the project model and the

rational design model.

3.1 The General Architecture of the

Proposed Approach

The architecture of our approach is, such that several

approaches in the literature, composed of three main

parts: an offline process, a base case and an online

process (Fig. 1).

• The Offline Process: this process starts from

modelisation (models + ontology) to the project

excavation. This part identifies and classifies

projects and the domain ontology.

• The Knowledge Base: it doesn’t only contain

instances of the ontology but also cases of

projects, project classes and problems arising

from the rational design.

• The Online Process: the online process is

from the acquisition of new project until

the project learning. the presents the

CBR reasoning cycle: which Development,

Remembering, Adaptation, enrichment,

validation and storage.

Figure 1: Architecture of the proposed approach.

Toward a Domain Ontology for Computer Projects Resolution: Project Memory Challenge

249

In this article, we will concentrate on the

modeling part of this approach in which we will

describe the three suggested models and the proposal

of domain ontology.

3.2 The Project Class Model

During the resolution of a computer project, we can

distinguish different classes in the same

organisation. Such as security, software engineering,

imaging, data base, artificial intelligence…

It is in this context, that we propose this model to

allow the leader to classify, from the beginning, the

project. This process can be done by specifying their

knowledge, its resolution method (scrum (Alain

Collignon, 2016), pert(Mahfouf, 2014)…), its

reasoning rule and its architecture. This model (Fig

2) is composed of three elements:

• Project Class: this element is composed of an

enumerative project list that belongs to this

class as well as a list of common

denominators (rules + keywords).

• Project Class Knowledge: All the knowledge

related to the project class in question are

associated to all the rules used in the

reasoning phase for this type of project class.

• Point of View: This component presents the

methods of conducting project class and the

type of used architecture.

Figure 2: The project class model.

3.3 The Project Model

The proposed project model (fig. 3) has three-

dimension. The choice of components of this model is

inspired from the composition of the project memory.

For this reason, in the following section we will

present a brief study of the notion of project memory.

3.3.1 The Memory Project: Definition and

Structure

According to (Nada Matta, 2014) a "project memory"

is a very limited part of a capitalization exercise of a

whole range of diverse experiences in the business.

This memory aims at facilitating the traceability and

the re-use of similar projects. It consists essentially of

two components which are the problem-solving

context and the method of resolution.

In (André, 2004) the "project memory" was

considered as a technique that approximates the

meeting often done at the end of the project because

it seeks to determine the same knowledge and

lessons learned during the project.

3.3.2 The Description of Proposed Model

The project model is composed of three elements:

• Project: This pillar gives general information

about the project. It includes the following

attributes.

• Project Name: gives the name, title or

project subject.

• Abstract: contains the objective, the

principle and the result of each project.

• Keywords: are essential words of a subject

or a project which allows them to be

identified.

• Project Team: This is the name of the person

in charge of running a project and managing

its progress (project manager, project-

director, supervisor, user, provider, project-

actor ...)

• Project Features: This component reflects all

the characteristics that a project can have

during its realization. Among these

characteristics we can quote the size, scope,

cost, time, complexity, type…

 Deliverable: this class is composed of two

sub-classes:

• Rational Design: This concept gives an idea

about the list of problems associated with

the solutions and suggestions given by the

leader of a project. In order to better explain

this component, we have proposed a model

which will be described in the following

subsection.

• Nature: The deliverable nature given by such

a project can be either: a service, document

or a product.

Figure 3: The project model.

KEOD 2018 - 10th International Conference on Knowledge Engineering and Ontology Development

250

3.4 Rational Design Model

This model (Fig 4) is presented using three essential

components:

• Problem list: each problem is described by its

name, its textual description and its attributes.

• Suggestion list: before reaching the final

solution the designers have proposed a set of

suggestion.

• Solution list: for each problem there is one or

more solutions that are defined (text) and

argued (arguements).

Figure 4: The rational design model.

After the presentation of these models we

noticed that there are a very large number of

different concepts which define the field of

computer projects such as types, characteristic,

deliverable...

4 DEVELOPING DOMAIN

ONTOLOGY

4.1 The Components of Ontology

Ontology can be defined as representative model

which presents the domain knowledge with explicit

specifications that feature interoperability between

human and machine,(Chun-che, 2015) it is

composed essentially of: classes, proprieties, axioms

instances and relation (Ahmed Maalel, 2011)

• Concepts: A concept can represent an object,

an idea or an abstract concept. They are also

called ontology classes in some works. A

concept can be divided into three parts: a term

(or several), a concept and a set of objects.

(Mejri Lassad, 2009).

• Relationships: Relationships reflect the

(relevant) associations between the concepts

presented in the analyzed segment of reality.

These relationships include the following

associations: (Subclasses of (generalization-

specialization) Part of (aggregation or

composition), Associates with, Instance of,

etc. (Chun-che, 2015).

• Properties: May include subproperties (and

superproperties). Ontologies define a set of

properties to be used in a specific knowledge

domain. There are many types of typologies of

properties such as the Inverse properties.

• Axioms of the ontology used to define the

semantics of terms (classes, relationships),

their properties and any constraints on their

interpretation. They are defined using well-

formed formulas of first order logic using the

predicates of ontology.

• Individuals: Instances constituting the

extensional definition of ontology; these

objects convey knowledge (static, factual)

about the domain of the problem (Ahmed

Maalel, 2011).

4.2 The Methodology of Ontology

Construction

The construction of ontology is a difficult task

requiring the implementation of elaborate processes

to extract the knowledge of a domain, manipulated

by computer systems and interpreted by human

being.

In literature there are many methods of

ontological construction. We present here our

choices of each step in the construction of domain

ontology for computer projects (Hernandez N. M.,

2007).

 The Text-To-Onto methodology is an

application for extracting ontologies from

corpora or web documents and it also allows

the reuse of existing ontologies (Marie-Noelle,

2009).

 The Onto Builder methodology, which allows

building ontology from web resources

(Ahmed Maalel, 2011).

 The METHONTOLOGY and KACTUS

(Hernandez N. M., 2007)which are designed

to be applied in more general settings. In

KACTUS, the methodology aims to reuse

existing ontologies and propose mechanisms

for this reuse. For METHONTOLOGY, it is

applied to clarify the various stages of

construction by respecting:

• Project management activities: planning,

quality assurance.

• Development activities: specification, conce-

ptualization, formalization, implementation,

maintenance.

Toward a Domain Ontology for Computer Projects Resolution: Project Memory Challenge

251

• Support activities (integration, evaluation,

documentation).

For this reason we have proposed a method

based on this methodology in order to construct our

domain ontology of the computer project domain.

4.3 Steps of Construction of the

Ontology

This methodology is offered through a set of steps, a

cycle of development of ontology that can be

adopted during the construction of a new ontology

(McGuinness, 2007)

a. Specification: The purpose of this step is to

provide a clear description of the problem

being studied and how to solve it. It clarifies

the purpose, scope and degree of granularity

of the ontology that will be constructed.

b. Conceptualization: During this stage, it is a

question of transforming the terms obtained

following the linguistic study of the corpus:

terms will be transformed into concepts and

the lexical relations in semantic relations. At

the end of this step, a conceptual model is

obtained. We distinguish in this phase the

two following tasks:

• The definition of concepts: it identifies the

concepts from the resources that were

originally specified in the specification phase.

• The hierarchy of concepts: it organizes the

concepts in a hierarchy that expresses the

subsumption relationship between concepts.

c. Formalization: The objective of this step is

to express, by means of a formal language,

the conceptual model obtained at the end of

the previous step (Ahmed Maalel, 2011).

This step makes it possible to add properties

to concepts, axioms, constrain the areas of a

relationship.

In other words, it is a question of defining

concepts according to a formal and extensional

semantics. It is also used formalize the relations that

exist between the concepts by defining their varieties

and the sets of extensions of concepts that they

connect (Hernandez N. M., 2007).

d. Implementation: This phase aims to move

the conceptual model to a model implement-

ed in one of the languages (OWL, OWL

Lite). For the implementation phase we work

with the “PROTEGE” tool (Ferdinand

Dhombres, 2010)

Figure 5: Proposed domain ontology.

KEOD 2018 - 10th International Conference on Knowledge Engineering and Ontology Development

252

This tool is the most popular and widely used

tool for ontology development (Naveen Malviya,

2011) It is a stand-alone open source platform that

provides a graphical environment for ontology

editing, visualization and control (constraint

checking) .It is the most popular ontology

publisher at the moment, serving as a reference for

a large community of users.

e. Maintenance: This phase can update the

ontology developed by adding,

modifying, or deleting concepts or other

elements of the ontology. The

Maintenance of ontology is very

important because it allows it to stay up to

date.

4.4 Main Classes of the Proposed

Domain Ontology (Fig 4)

The field of the "computer project" includes a large

number of concepts related to the project concept),

such as the identifiers of a project, the

characteristics and the types of deliverables...

In the next section we will describe the main

concepts of our project, proposed ontology.

• Project: This is the main class of our

domain ontology. A project is defined by a

set of attributes that are the project name,

summary, keywords, project manager, and

project team.

• Characteristic: a project has a set of

characteristics that can be differentiated

from one project to another: Scope,

complexity, size, delay and cost, resource.

Each of them can have subclasses

(complexity: complex, simple, innovation ..,

size: small, large ...,).

The concept «scope feature, can be either a

professional project (business project), a

research project (web, security, data base,

imaging, networks ...) or a study project

(license, master's degree or thesis.)

• Deliverables: each project is characterized

by a return value, this value can be of three

types:

• Product: A product can be hardware or

software

• Service: service offered online or offline.

• Document: It can be a site or a text. This

text can be a report (design report, usage

report or technical report).

Figure 6: Creation of an individual and instantiation of ontology.

Toward a Domain Ontology for Computer Projects Resolution: Project Memory Challenge

253

4.5 Creation of an Individual

We have suggested this ontology in the hope of

creating a strong relation between this notion and the

case-Based reasoning system (CBR) and this can be

done by different factors:

• Ontologies play an crucial role in CBR

systems because they can reduce the effort to

acquire

• Knowledge in the different stages of reasoning

(Chun-che, 2015).

• Ontologies are effective ways to formalize

structure, store and used knowledge.

The final instantiation of this ontology

(individuals + instances) is actually the new case on

which our reasoning is based (fig 6). They help to

establish a common vocabulary to describe the case,

or the model knowledge needed to index and

organize the event.

We have advanced our thesis research topic to

instantiate our proposed ontology.

5 CONCLUSIONS

Through what we have presented in this paper, it

turns out that the notion of ontology represents a

very effective approach to introduce knowledge.

Throughout this paper, we have tried to clarify

the notion of ontology by presenting certain

definitions from their types and their components.

In addition, we have described several

methodologies in the construction of ontology and

we have finished this section by proposing a method

of construction of a domain ontology based on

METHONTOLOGY methodology.

The domain ontology explains the concepts and

relationships that can be found in the field of

computer projects, this one can be extended with

task ontology.

In the near future works we will focus on

interrogate our ontology either by using the

SPARQL-QUERY (“Protégé’s tool) or by proposing

a short algorithm. Then we will tackle the producing

of our knowledge base which will be built from the

data set existing in the “Hall”.

REFERENCES

Ahmed Maalel, M. L. (2011). Toward A Knowledge

Management Approach based on an Ontology and

Case-Based Reasoning (CBR) Application to Railroad

Accidents.

Alain Collignon, J. S. (2016). Agile management

methodology of a project (scrum). https://www.cairn.

info/revue-i2d-information-donnees-etdocuments-2016

-2-page-12.htm .

André. (2004). Organizational memory breaks with

individual memory ", Communication and

organization. La france.

Bekhti, S. (2003). Dypkm a dynamic Process for Definting

and reusing Project Memory/human Machine

Interface. .

Benoît Eynard, M. L. (2001). Construction of a project

memory in mechanical engineering using web

technologies. cairn info , 5 (155-147).

Chun-che, L. w. (2015). A rough set based corporate

memory for the case of ecosystem”. Taiwane.

médecine prénatale. France.

Ermine. (2001). protect corporate memory, communication

in the context of cyclede meetings technologies key.

Ferdinand Dhombres, J.-M. J.-C. ( 2010). Methodological

choices for the construction of a domain ontology in

prenatal medicine.

Harani, Y. (1997). A multi-model approach for the

capitalization of knowledge in the field of design.

These of the INPG, specialty in Industrial

Engineering.

Hernandez, N. M. (2007). Modeling context through

domain ontologies. Springer Science + Business

Media .

Mahfouf, N. (2014). Planification et Ordonnancement

d'un projet a moyens limités au sein de L'ENGTP.

http://dlibrary.univ-boumerdes.dz:8080/handle/123456

789/4069.

Marie-Noelle, E. K. (2009). Extraction of Terms,

Recognition and Labeling of Relationships in a

Thesaurus - Towards an Ontology. Poplawski.

McGuinness, N. F. (2007). Ontology Development 101: A

Guide to Creating Your. , Stanford, CA, 94305:

Stanford University.

Mejri Lassad, H. H. (2009, 09 06). a unified generic model

of problem representation and resolution for reuse of

knowledge. pp. 132-148.

Nada Matta, R. A. (2014). “Hybrid System fr

Collaborative Knowledge Traceability "an Application

to Bussiness Emails”.

Naveen Malviya, N. M. (2011). Developing University

Ontology using protégé OWL Tool: Process and

Reasoning. International Journal of Scientific &

Engineering Research Volume 2, Issue 9.

Paula Potes Ruiz, D. N. (2012). Reasonably collaborative

from cases in. 9th International Conference on

Modeling, Optimization &.

Serrano, L. (2014). Towards user-oriented knowledge

capitalization: Automatic extraction and structuring of

information from open sources. france: HAL Id: tel-

01082975.

Zacklad, M. L. (2014). MEMO-net, a groupware using the

DIPA problem solving method for capitalization and

knowledge management in design projects. Troyes

Cedex.

KEOD 2018 - 10th International Conference on Knowledge Engineering and Ontology Development

254