A Generic Model Of Corporate Memory: Application To Industrial Systems
Lotfi Admane, Institut National d’Informatique,
Algeria
ABSTRACT:
This paper presents an
industrial system model and a model of corporate memory supporting it. The
industrial system model decomposes the industrial system through a set of
components. The corporate memory model views the latter through two types of knowledge : 1) Skill Knowledge, which constitutes the main
knowledge capital of the company, refers to its basic skill and 2) Theme
Knowledge representing a specialized knowledge or a knowledge related to a
given field.
The objective of the
corporate memory model, named ReCaRo, is to
capitalize knowledge to its systematic re-use. ReCaRo
builds corporate memories which have a multi memory
architecture. This architecture means that every memory will be made up of five
communicating ones. The modelling using ReCaRo will be done in two steps :
1) modelling the industrial system and 2) implementing the re-use principle in that
system.
1. Introduction
This work was motivated by
the observation of the strong tendency of the company to be specialized. To be competitive,
the company standardizes its processes and its resources. It, often, handles
the same entities for different actions. These entities can be physical
objects, rules, processes, etc. Progressively and using these entities, the
company constitutes, in the long term, a knowledge capital. However, this
knowledge capital is often scattered on the experts of the company and in
documents. It is very volatile. The objective is to collect it, organize it and
preserve it for re-use purposes. This preservation is done through the concept
of corporate memory (Brooking, 2000; Kingston & Macintosh, 2000; Ackerman, 1998;
Abecker et al, 1998; Pomian,
1996; Vanheijst et al, 1996). This capital is then
re-used in different situations in order to reduce the costs and the time of
development. It is the concept of the knowledge creating company (Nonaka & Takeuchi, 2002) and of working knowledge (Davenport
& Prusac, 2000).
In this paper, we propose a
corporate model and a generic corporate memory model supporting it. The
corporate memory model is named ReCaRo which is the
acronym of REsource, CAse
and ROle which are the main concepts of the memory. In
this work, we emphasize, particularly, on the re-use problem of the knowledge
contained in the corporate memory.
To build the corporate
memory, we propose an approach in two stages. In the first stage, we will
propose a modelling of the company which supports the development of corporate
memories allowing the easy and systematic re-use of knowledge capital. In this
research, we are interested in the industrial companies which we will call industrial
systems. In a second stage, we will propose a generic corporate memory
deduced from the industrial system.
This paper is divided into four
main parts. The first one presents the concept of knowledge and its reuse. The second
one presents the model of the industrial system. The third presents the
proposed generic model of the corporate memory and, finally, the last one
presents the application of the ReCaRo model for the
conception of a corporate memory in the design of industrial systems for liquid
hydrocarbons transport.
2. Corporate Memory And Reuse Of Knowledge
There are many
classifications for corporate knowledge. In this section, we give some
theoretical elements regarding the knowledge management in the company as well
as the concept of knowledge re-use. We will not try to define the corporate
knowledge, but we will position ourselves directly within a framework of
corporate knowledge management while trying to determine its main
characteristics. Among these characteristics, we have the usual distinction
between data, information, comprehension and knowledge (Boersma
& Stegwee, 1996). The aspects of contextualization
of the knowledge were particularly developed in sociology (Tang-Ho & Lamontagne, 2000) and in cognitive psychology (Poitou, 1997; Pomian, 1996) introduced
the distinction between "knowledge", "information" and
"data". Ackerman & Halverson (1999) introduced the distinction
between know-how and the knowledge. Danveport & Prusac (2000) evoked the link between knowledge and context.
Finally Ermine et al (1998) introduced the concept of knowledge system. The
definition of the concept of knowledge in the organization remains very prone
to discussion. As far as we are concerned, we remain in the optics of Davenport
& Al Prusac (2000) which are interested in the
problem of knowledge re-use.
There are
several classifications of corporate knowledge. For example, the typology of Buckingham-Shum
(1997) classifies knowledge according to its specialization. The typology of Brooking
(2000) and Pomian (1996) classifies knowledge
according to its type.
An important point in this paper
is the re-use of this knowledge. In the most general case, "To
re-use" means, to use again existing elements. In the case of knowledge
management, the term re-use means, to use one or
several existing resource components in order to create new components within a
minimum search time and adaptations. They have to be lower than those necessary
to the construction of new components offering the same functionalities (Zdrahal, 1998). Most works on the principle of re-use introduce
the concept of reusable component (Bushmann et al, 1998).
These works introduce the reusable component as being an object of the
organization described through a set of characteristics, often descriptive. For simplification needs, these reusable components are
often gathered in classes.
3. The
Industrial System Model
In this section, we will answer the
following question: how can one model an industrial system from a point of view of
a corporate memory?
According to Sherif
& Mandviwalla (2000), the modelization
of industrial systems is faced to the diversity of these latters.
Mohanty & Deshmukh (1999)
proposes a strategy for the modelization of
industrial systems that takes into account this difficulty.
For our model of the industrial
system, we have exploited the idea developed in projects DECIDE (Alquier, 1997) and PRIMA (Alquier,
2000). These projects propose a classification of the reusable components for
cost management in the design of new products. The model of the company that we
propose views the company through the entities it re-uses. Any industrial
system, therefore, will be modelled as a system constituted of, or handling,
two types of components:
Ø The skill
components: They represent every physical or
logical object which constitutes the basic skill of the company.
Ø The theme
components: They represent all that is necessary
to operate the system, such as the consumed resources, the inputs and the
outputs of the system etc.
The two components’ types are
in interaction and are connected according to a suitable logic. These
components constitute the main objects on which a knowledge capital is
developed. The corporate memory model that we propose aims to capitalize this
knowledge with a re-use aim.
4. The Corporate Memory
Model
The corporate memory model that we
propose as a support to the model of the industrial system
handles two types of knowledge and has a multi-memory architecture. In this
model, every memory is built around five main memories that we present below:
Ø
Reusable resources’ memory (R. R. Memory): In the first stage, it is necessary
to make an inventory of the corporate knowledge capital. We propose to organise
the knowledge capital into two types. The one relative to skill components that
we will call the Skill Knowledge and that relative to Theme Components that we
will call the theme knowledge. This is done through an inventory of the industrial system’s components. Each
component will be listed in the form of what we will call the reusable
resource and will constitute the memory of reusable resources. Each
reusable resource answers a well defined model. To conceive the memory of the
reusable resources is, finally, to make an inventory of the various types of
components and modelling each one of these types.
Ø Roles’ memory: In a second stage, it is necessary
to build the memory of the roles. A role describes an element of the reusable
resource context. The objective of the roles is to ensure the connection of a
resource’s use to its context of use. Very often, modelling a role memory
means making an inventory of the whole set of roles that describes and comments
the knowledge capital, and proposing a model for each type.
Ø
Cases’ memory: A case of use represents the
description of the use or the re-use of a reusable resource in a given context.
It is defined by a reusable resource, to which it was decided to add a set of
roles. Each role is carrying a single semantic which relates to the described
part of the context. In a third stage, it is necessary to make an inventory of
all the experiences, around these reusable resources. Each experience is
represented by a case of use. To constitute the memory of the cases, it
is necessary to make an inventory of all the types of cases. Each type of case
will be represented by a model.
Ø Networks of cases’ memory: A network of cases represents
the description of the coordinated use of several cases for the realization of
a common and single goal. It is a set of several cases connected by roles. In
this situation, the roles are considered as connectors carrying a semantic.
This memory is used to describe experiences which are too complex to be
described by cases.
Ø Contexts’ memory: The context is a description of the
situation in which the case was carried out. The concept of context is
complementary to that of a role. It is used to describe situations which are either
too rare or too complex to be described by roles.
According to this, every corporate
memory will be composed by those five memories. These memories are connected
according to the logic described by Figure 1. The cases' memory consists of the
connection of the reusable resources, the memory of roles and the memory of
contexts. The memory of the cases network consists of the connection of the
cases' memory, the roles’ memory and the contexts’ memory. The
global design of the corporate memory is thus summarized as follows: make, for
each of the five concepts, an inventory of its various forms and propose for
each form an adequate model.
Figure 1: General Architecture Of The Corporate Memory
5. Application Of The Model To The SH-TRC Project
The aim of this section is
to validate the proposed model on a real case of corporate memories' modelling.
We will present the SH-TRC project as well as the built corporate memory.
This project aims to set up
a corporate memory of all the capital experience and knowledge acquired during
the design, the renovation, the maintenance or the extension of the
transportation of liquid hydrocarbons. Each one of these actions is done through
a study which gives rise to a specification, describing, for example, the
requirements and the technical features of the future industrial system. This
specification constitutes one of the main sources of knowledge.
The goal of this project is
to help the designers of future industrial systems to carry out their tasks by
assisting them in the design phase. This assistance will be done by proposing
to the designers a set of components, resulting from the memory of components,
and approaching the most their needs.
5.2. The Industrial Model Of The SH-TRC Project
As described in section 3, the industrial system of the SH-TRC project
is viewed through two types of components:
Ø
Skill components - we have listed two types of skill components. The component of type
product which represents any physical element entering in the
composition of an industrial facility (for example: pump, circuit breaker, pipe etc.) and the
component of type process which represents all the dynamics of the
industrial system.
Ø
Theme components - Among the themes studied in the project SH-TRC,
the topic retained is the conception of
the specification draft. The development of the latter offers to the users an assistance in the specification of future installations. In
this theme, we have found two types of components: the component of type portion
of text and the component of type graphic element. The component of
type portion of text gathers all that was written around a skill component and
the component of type graphic element gathers all the diagrams and graphs
associated to a component.
5.3. Architecture Of The Proposed Memory In The SH-TRC Project
In accordance with the
generic architecture proposed in section 4, the memory of the SH-TRC project
will be composed by five memories that we will present below:
5.3.1. The Memory Of Reusable Resources
This memory gathers the
four types of reusable resources: product type, portion of text
type, graphic element type and process type. In this paper, we only
describe the product type, the portion of text type and the graphic
element type. The reusable resource of process type is described in Admane et al (2004).
Memory
of the reusable resources of type product: It is the main resource among the ones of the skill
type. Almost all the corporate memory requests are on it. The product is regarded
as the element of the finest granularity. It can be used in the composition of
an industrial facility, or in the composition of another product. The model of Figure
2 describes the product isolated from any use. The structural properties
describe the resource. The properties of environment describe the interaction of
the resource with its environment. The properties of re-use describe the
possible re-use forms of the resource. Each instance of a component's model
will become an element of the memory of the reusable resources of type product.
Figure 2: Model Of Reusable Resource Type Product
The
example of Table 1 is an element of this memory. We represent this component as
a record. This example describes a programmable pump. The characteristics of
the re-use inform that this pump can be used as an amplifier of flow. It can
also be used in manual or automatic mode.
|
Reusable Resource Product: pump P217 |
|
|
Name: |
ZPHMP2002-1 |
|
Description: |
Programmable Hydraulic pump |
|
Standard
Properties: |
Double pumping, … |
|
Standard
Values |
Max Power: 100 bars, diameter of entry:
300 mm, … |
|
Energy: |
Diesel |
|
Climatic zone: |
Arid, tropical |
|
Constraints of reuse: |
Product not corrosive |
|
Technical
Family: |
Pumping, amplification
flow … |
|
Cost
Family : |
High |
|
Reference
value: |
|
|
Form
of reuse: |
Amplifier of flow |
|
Properties of reuse: |
Flow parameters, modification section … |
|
Values
of reuse: |
|
|
Constraints of
reuse: |
|
|
Possible
State: |
Automatic, manual … |
Table 1: Example Of
A Reusable Resource Of Type Product
Memory of
reusable resources of type portion of text: It is a resource of the theme resources’
memory. This resource represents any portion of text that seems interesting. It
is described in Figure 3. The re-use properties of type nature describe the
form or the length of the text and those of the re-use type of usage give the
type of the text (descriptive, modifying, etc.)
Figure 3: Model Of
Reusable Resource Of Type PORTION OF
TEXT
The
example of Table 2 describes a reusable resource of type potion of text.
It is an assembly notice which is a short technical text that describes a
process. The text itself is in the Content property.
|
Reusable resource PORTION of TEXT : note of assembly |
|
|
Name : |
Notice 111 |
|
Description
: |
Note of assembly of the programmable
hydraulic pump |
|
Content : |
<Text > |
|
Writer
: |
Name of the writer |
|
Text
Form : |
Technical Text |
|
Length
of text : |
Short |
|
Type
of text : |
Process |
Table 2: Example Of
A Reusable Resource Of Type PORTION OF TEXT
Memory
of reusable resources of type graphic element: The reusable resource of type graphic element
is, also, a resource of the theme resources memory. This resource represents
any graphic element extracted from the specification which seems interesting. It
is described in Figure 4. The properties of re-use of the type nature describe
the form or the dimensions of the graph and those of type of usage give the
type of the graph (descriptive, modifying, assembly, etc). A reusable resource
of graphic type can be modelled exactly like a reusable resource of textual
type.
Figure 4: Model Of
Reusable Resource Of Type graphiC element
The roles serve to describe
all or a part of a reusable resource within a particular use. All the semantics
carried by the role relates to the evoked reusable resource. The model of roles
is described in Figure 5.
Figure 5: Role’s Model
In the SH-TRC project, we proposed three models of roles:
Model
of internal roles: These
roles implement only one reusable resource: the current resource. The global model of this role is described
in figure 7. It is described by a set of properties which are: the name which
is used as an identifier and a short description of the role. As an example of
internal roles, we can inventory all the roles which aim is to describe the invoked
reusable resource. Among these roles, we find the role Mechanical which describes the reusable resource as being a
mechanical component, the role Manual_mode which means that the component is used in a
manual mode, etc...
Model
of role of connection: The roles of connection are used to describe all
the interaction that the invoked reusable resource can have with another
reusable resource. This latter is called dependent
reusable resource. The general model of this role is identical to the
internal role except that it implements two reusable resources. These roles can
be for example: connected_to
which informs that the current resource was connected to another, adapted_on which means that the component can
be adapted on another component or commented_on which means that the component is commented
on by another resource (of type portion of text or person that we can define).
Model
of role of mediation:
The roles of mediation type are used to describe the way in which the invoked
reusable resource comes between two dependent reusable resources. This resource
is used as a mediator between the two dependent reusable resources. The role of
mediation, thus, involves three reusable resources. We give, as an example, the
following roles: To assemble
which means that the reusable resource is used to assemble two other reusable
resources; to annotate which
means that the reusable resource (of textual type) is used to put a comment on
a portion of text associated to a resource etc.
Table 3 presents some
elements of the roles’ memory.
|
Types |
Roles |
|
Internal Roles |
Physical Obj, obj Informational, machine, module, hydraulic,
mechanics, electric, location, make modification, section reduction, descriptive
text, operating mode Position-function, Modify- characteristic, suppress -component,
(joint), etc. |
|
Connection’s Roles |
Adapt on, connected to, connectable with, non compatible with, used in, described by, describe, schematized by,
schematize, Add-component, Obligatory composition, Optional Composition, Specialization, Generalization, Reuse, Derivation, Equivalence, Obligatory need, Optional need, Induction, precede, following, before, after, etc. |
|
Mediation Roles |
To assemble, to adapt, annotate, informs, etc. |
Table 3: Extraction From The Roles’ Memory
5.3.3. Memory Of
Contexts
We propose models of contexts
which make it possible to describe situations in the form of texts. Practically,
we built a single model of context. This model is illustrated in Figure 6. It represents the context as being a portion
of text described by the context code, the context name, and the family of
context. We define for example the contexts:
Ø
Desert: knowledge is valid for desert regions
Ø
Renovation: knowledge is valid in a situation of renovation
Ø
Reduction of section: which means that the described knowledge is valid for
problems of conducts section reduction.
Figure 6: Model Of Context
5.3.4. Memory Of Cases
The aim of these models is to represent the cases of use of reusable
resources. Each model describes a type of a well defined case. A model of cases
is composed of the triplet: reusable resource, role and context. It is
described by Figure 7.
Figure 7: Model Of CASES
Structural properties: (name, description) describe the
case itself. The reusable resource represents the resource implemented in this
particular case. It results from the memory of the reusable resources to be
added as a case in the memory of the cases. The roles are used to document the
way in which the reusable resource was
used in this case. The role is invoked by its name. The context describes the
context in which the case was built. The action explains the action defined by
the role on the reusable resource. The example of Table 4 describes a case of
use of a programmable pump like a regulator of pressure.
|
Case: amplification of pressure |
|
|
Description: |
<Description of the
case> |
|
Reuse Res.: |
pump P217 |
|
Role: |
Position-function, Modify- characteristic, Add-component (joint) |
|
Context: |
Increase power |
|
Actions: |
Position the pump in automatic mode, increase its
section, remove the joint of origin and
replace it by a hermetic one. |
Table 4: Example Of A Reusable Resource Of Type
CASE
5.3.5. Memory Of Scenarios Of Cases
There are, in reality, situations which are meaningful only after the
description of many cases of use of several reusable resources. For these
situations, we introduce the concept of scenarios of cases. The latter are
defined like the description of the interaction of two or several cases for the
realization of a common objective. The general model of the scenario of cases
is defined in Figure 8. The cases are connected by sequence or logic roles (and, or, oux
etc).
Figure 8: Model Of A Scenario Of Cases
For example, the connection
of two conducts cannot be described by a case. It is the combination of four
cases of use. The built scenario can be schematized as in Figure 9. In this
example, Case 1 means the assembly of a core of reduction (to reduce the
diameter of a conduct), Case 2 means assembly of a core of increase (to
increase the diameter of a control), and Case 3 means assembly of a ring
of connection (connection of two conducts) and Case 4 means regulating
assembly of pressure (regulation of the pressure)
Figure 9: Example Of
A Scenario Of Cases
6. Conclusion
In this paper we presented
a model and a methodological way for designing corporate memories based on the
re-use principle. Two ideas were developed:
Ø
The
first one relates to the architecture of the corporate memory. We proposed multi-memories
architecture. This means that every memory developed according to the ReCaRo model will be composed of five communicating
memories: the memory of the reusable resources, the memory of the roles, the
memory of the cases and the networks of cases and the memory of the contexts.
Ø
The
second idea relates to the implementation on industrial cases. The main problem
that we had to solve was the definition of the concept of reusable resource in
the field of the hydrocarbon transport. For that, we proposed a modelling of
the industrial system through two classes of components: skill knowledge and
theme knowledge.
The memory of corporate
knowledge offers to the technicians all the help with the industrial systems
design. The connection of the corporate memory to the documentary theme memory
offers them the assistance with specification when designing new installations.
We decided to implement the corporate memory as a data base. The set of models
and reusable resources was implemented as a set of data bases. Admane et al (2002) give all the details for this modelling.
To capitalize, in the long term
this work, the idea is to develop mechanisms to collect produced knowledge. In
our case, this knowledge is:
Ø
Generic
models of reusable resources: they are collected throughout the dissemination
of the suggested method. These models are standardized, and given to the
designers of corporate memories as reusable generic models;
Ø
Generic
models of reusable roles: they are collected in the same way as the reusable
models of resources. The models of roles are standardized and classified;
Ø
Listing
of the most usual roles: one could collect the roles themselves because they
can be reusable in their state. Their capitalization becomes making an
inventory of those roles, organizing them and
proposing them to the users.
7. References
Abecker, A., Bernardi, A., Hinkelmann, K., Kuhn, O. and Sintek,
M. (1998). Toward a technology for organizational memories, IEEE Intelligent Systems,
May/June 1998, p40-45.
Ackerman,
M. and Halverson, C. (1999). Considering organizational memory: Processes,
Boundary Objects, and Trajectories, Proceedings of the 32nd Hawaii
International Conference on System Sciences, 1999.
Ackerman, M.S. (1998);
Augmenting organizational memory: a field study of
Admane, L., Benferhat, R. Boukraa, D., Ghomari, A., Bouchama, S., Terr S., Alquier, A.M. & Tignol, M. H.
(2002). Un modèle de projet
et une mémoire de projet pour la conception de cahier de charges. Proceedings of 7th MCSEAI (pp103-110).
Admane, L., Benferhat,
R., Ghomari, A. & Terr,
S. (2004). Modélisation de projets de mémoire
d’entreprises. Proceedings
of first information systems optimisation international congress, (pp67-75).
Alquier, A. M. (2000). Project
Risk Management. www.esi2.us.es/prima.
Alquier, A. M.(1997).
DECIDE : DECIon Support for
Optimal Bidding in a Competitive Business Environment, projet
n°22298. CEE projets, Esprit
program.
Boersma, J. & Stegwee,
R. (1996). Exploring the issues of Knowledge Management. Proceedings of
the Information Resources Management Association International Conference (pp
127-135).
Brooking, A. (2000).
Corporate memory: strategies for knowledge management.
Buckingham-Shum,
S. (1997).
Negotiating the Construction and Reconstruction of Organisational Memories,
Journal of Universal Computer Science (Special Issue on IT for Knowledge
Management), Vol. 3, No. 8, p899-928
Bushmann, F., Meunier, R., Ronhert, H., Sommerland, P. &
Stal, M. (1998). Pattern oriented software architecture: a
system of pattern. Johw Willey and Sons Ed.
Ermine, J-L , Barthelme, F., and Rosenthal-Sabroux, C. (1998). Architecture for
knowledge evolution in organizations, European Journal of Operational Research.
Vol. 109, No. 2, September, 1998, p414-427.
Mohanty, R.P. and Deshmukh,
S.G. (1999). Evaluating manufacturing strategy for a learning organization: a
case, International Journal of Operations & Production Management, Vol. 19,
No. 3-4, 1999, p308-327.
Nonaka, I.
& Takeuchi,
H. (2002). The Knowledge-creating Company: How Japanese Companies Create
the Dynamics of Innovation. , Ed. Hardcover.
Pomian, J. (1996). Mémoire d'entreprise :
techniques et outils pour la transmission du savoir.
Sherif, K. and Mandviwalla,
M. (2000).
Barriers to actualizing organizational memories: Lessons from industry,
Proceedings of the 33rd Annual Hawaii International Conference on System
Sciences, 2000, p68.
Tang-Ho,
L. and Lamontagne, L. (2000). Structuring
Organizational Memories using Multi- Dimensional Knowledge Networks,
Proceedings of the ECAI'2000 Workshop on Knowledge Management and
Organizational Memories, 2000.
VanHeijst, G., VanDerSpek, R,
& Kruisinga, E. (1996). “Organizing
corporate memories”. In Proceedings of workshop
on Knowledge Acquisition for Knowledge-Based Systems (pp. 87-96).
Zdrahal, Z. (1998). An
Intelligent Brokering Service for Knowledge-Component Reuse on the
World-Wide-Web, In Proceedings of the 11th Workshop on Knowledge Acquisition, Modeling and Management (KAW«98).
Contact the Author:
Lotfi ADMANE, Institut
National d’Informatique, BP 68 M, Oued Smar, 16270, Algiers,
Algeria; Tel: 00 213 70349835; Fax: 00 213 21912186; Email: l_admane@ini.dz