INTEGRATION OF KNOWLEDGE MANAGEMENT IN
PRODUCT DESIGN PROCESS
Yang Xu and Alain Bernard
IRCCyN, Ecole Centrale de Nantes, 1 Rue de la Noë, 44321 Nantes, France
Keywords: Knowledge management, Product development process, Modelling.
Abstract: In order to keep up with the high adaptability of the product design requirements, it is essential to integrate
knowledge management into product design processes. This paper proposes a model that conjoins the
objects, functions, constraints and knowledge, analyzes the task evolution process by means of state
characterization. By using the schema based on the model, it is practical and efficient to find appropriate
knowledge to a given general object in the context of certain constraints.
1 INTRODUCTION
In this era characterized by worldwide competition,
knowledge management is playing a key role in
every aspect of our lives, especially in enterprise
productions (Bernard and Tichkiewitch, 2008). In
order to improve the possibility of success, a variety
of knowledge management (KM) and computer
integrated manufacturing (CIM) models have
emerged in recent years, such as the FBS-PPRE
model (Labrouse, 2004), advanced technologies
based on CIM (Nagalingam and Lin, 2008), the
GRAI integrated method (GIM) (Doumeingts et al.,
2000), the object-oriented approach based on the
UML (Unified Modelling Language) modelling
(Merlo and Girard, 2004), etc. They all have
particular contributions to the modelling and
analysis of the knowledge management.
Although the notion of knowledge management
has different explanations and its research content is
fairly abundant, the aim of giving appropriate
knowledge to appropriate persons at appropriate
times is widely accepted. Therefore, one of the
challenges for the knowledge management is to
clarify the expectations with constraints in order to
propose the corresponding knowledge to satisfy the
need.
This paper focuses on the architecture and the
process to match the objects, functions, constraints
and knowledge, and proposes an OFCK model
applied to the product design process.
The paper is organized in four parts. Besides the
introduction section, Section 2 proposes a model that
conjoins the objects, functions, constraints and
knowledge, Section 3 analyzes how knowledge is
integrated into the product design process, and
Section 4 concludes the paper with several
perspectives.
2 THE OFCK MODEL
From the experience in product development process,
we know in order to accomplish a complex task, it is
appropriate to divide it into sub-tasks and sub-sub-
tasks until the objects of the indivisible tasks can be
reached easily. The relationship among objects,
constraints and knowledge should be well studied in
order to find appropriate supports effectively.
Consequently, an Object-Function-Constraint-
Knowledge (OFCK) model is proposed to abstract
the relationship among these four aspects in
knowledge management.
2.1 Objects, Functions and Constraints
In practice, any task is consisted of two parts: the
object and constraint, as nothing can be done
without any constraint. Thus, a task is divided into
two sets: the set of objects and the set of constraints,
noted as follow.
• The set of objects
12
{ , ,..., }
op
Sooo= , where
12
, ,...,
p
oo o are called atomic objects. Each
338
Xu Y. and Bernard A..
INTEGRATION OF KNOWLEDGE MANAGEMENT IN PRODUCT DESIGN PROCESS.
DOI: 10.5220/0003093003380341
In Proceedings of the International Conference on Knowledge Management and Information Sharing (KMIS-2010), pages 338-341
ISBN: 978-989-8425-30-0
Copyright
c
 2010 SCITEPRESS (Science and Technology Publications, Lda.)

atomic object
i
o corresponds to an atomic
function
i
f , in other words,
i
f can achieve
i
o and they are one-one corresponding:
ii
of⇔ . All these
i
f form the set of
functions
12
{ , ,..., }
f
p
Sfff= .
• The set of constraints
12
{ , ,..., }
cq
Sccc=
,
where
12
, ,...,
q
cc c are called atomic
constraints.
According to the idea of divide-and-conquer that
partitions the problem into fine scale sub-problems
and assembles of the solutions of the partitioned
sub-problems to constitute the overall (Gravemeier
and Wall, 2008), a task
T
can be divided into
several atomic sub-tasks that can be solved simply,
thus, a set of atomic tasks is noted
as
},...,,{
21 pt
tttS = , where
i
t is the atomic sub-
task. Each
i
t is an ordered pair consisted of an
atomic object and a set of constraints, i.e.
}),...,,{,(
21
i
p
ii
ii
cccot =
. Due to
ii
fo ⇔ , we
have:

}),...,,{,(}),...,,{,(
21
*
21
i
p
ii
ii
i
p
ii
ii
cccftcccot =⇔=
As a result, the task
T
is decomposed into a set
of ordered pair
}),...,,{,(
21
i
p
ii
i
cccf
.
2.2 The Knowledge
Knowledge should be organized properly in order to
be useful in practice, so a knowledge base KB is
introduced as the set of the organized knowledge
and it is defined as follows.
Definition 1. The knowledge base KB is a group of
vectors, noted as:

⎪
⎪
⎩
⎪
⎪
⎨
⎧
=
=
=
⎪
⎪
⎭
⎪
⎪
⎬
⎫
⎪
⎪
⎩
⎪
⎪
⎨
⎧
=
},...,,{
...
},...,,{
},...,,{
,
...
,
,
21
22
2
2
12
11
2
1
11
2
1
2
1
n
m
nn
n
m
m
n
n
kkkK
kkkK
kkkK
where
K
K
K
KB
In this definition,
j
K is the knowledge unit,
j
i
k is
the atomic knowledge, and
j
j
i
Kk ∈
.
2.3 The
Function-Constraint-Knowledge
Map
One of the most important targets in knowledge
management is to find required knowledge in the
context of the given objects and constraints, in other
words, to map the objects, constraints and
knowledge in a proper way. As
ii
fo ⇔
, the aim
becomes to build a Function-Constraint-Knowledge
(FCK) map.
Hypothesis 1. An atomic knowledge
j
i
k
possesses
an atomic function
y
f in the context of a certain set
of constraints
},...,,{
21 x
ccc , which is noted as:
{
}
yx
j
i
fccck ⎯→⎯⊗
Δ
,...,,
21
This hypothesis is based on the facts that the
knowledge is regarded as a kind of support in
solving practical problems, so it must have a kind of
function, otherwise, can it be called “knowledge”?
As a result, in the FCK map, there are groups of
“
k
,
f
, },...,,{
21 x
ccc ”.
2.4 Knowledge Seeking
From the previous analysis, once an atomic task
i
t
is given with related constraints, required atomic
knowledge can be found through the FCK map.
Then, an effective method should be referred to in
order to seek the knowledge in the KB that can
provide the atomic knowledge required. One idea is
to apply the inverted list used in the practical
information retrieval systems to support text
searching (Lester et al., 2006), which is an effective
structure that maps from a query term, the atomic
knowledge
1
k , to a posting list that identifies the
knowledge units that contain that term.
3 KNOWLEDGE INTEGRATION
IN PRODUCT DESIGN
PROCESSE
The process of product design is one of the most
complex stages in production which need both
explicit and tacit knowledge. For example, as simple
as a chair, its design process is not as easy as we
thought. Concisely, a designer has to consider about
INTEGRATION OF KNOWLEDGE MANAGEMENT IN PRODUCT DESIGN PROCESS
339

Figure 2: The application process of the OFCK model.
its comfort, mechanical support, recycle possibilities
and manufacturing cost. The complicated
relationships of function, constraints and knowledge
are involved. For example, a designer may use
carton as the main material of the chair to satisfy the
task of “the chair should be recycled”, but can it
meet the requirement of “to support a man of 80kg”?
To achieve this object, different chair architectures
are possible according to different chair materials,
which are related to cost consideration. At the same
time, chair architecture should also take the task of
aesthetic design into account. In a word, design
processes are very complicated, so we need a
knowledge management model to simplify and
realize them.
One basic idea is to decomposed a whole task as
sub-tasks and arrange these sub-tasks. The OFCK
model described in the section previously presented
is atom-unit oriented, so a general task should be
decomposed in order to adapt to the model. To begin
with, the product development process is
characterized as a basis of task decomposition.
3.1 Characterization of the Product
Development Process
In order to characterize the product development
process, a finite state machine of product (FSMP) is
proposed and defined as follows.
Definition 2. A finite state machine of product
(FSMP) is a hextuple
FEQ ,,,,,
0
δ Π∑
, where:
Q
is a finite and non-empty set of product
states;
∑
is a finite and non-empty set of supports
(knowledge, manipulations, etc.);
Π is a finite and non-empty set of constraints;
δ
is the state transition function:
QQ →Π×∑×:
δ
;
0
E is an initial product state and QE ∈
0
;
F
is the set of final states and
QF ⊂
, and
there is at least a state FE
n
∈
.
In fact, the aim of the OFCK model is to find
appropriate knowledge to realize the supports
according to the given task and constraints.
3.2 Task Decomposition
In order to achieve the main object, the general task
is decomposed by splitting the whole project into
sub-parts, and the process of achieving the general
task is regarded as an evolution process of the
product states, shown in Figure 1.
Figure 1: A product development process described by
FSMP.
In Figure 1,
i
E is the product state and
i
t is the
atomic task to change the related state. Inferred from
the OFCK model, knowledge seeking is processed
as:
i
mapFCKthe
i
p
ii
ii
i
p
ii
ii
k
cccftcccot
⎯
⎯
⎯
⎯
→
⎯
=⇒=
21
*
21
}),...,,{,(}),...,,{,(
Consequently, the formalized definition the
general task
T
is given as follows.
Definition 3. The general task
T
is what to be
accomplished to realize a product development
process, in other words, to evolve the product
P
from
0
E to
n
E . The product
P
is a finite state
KMIS 2010 - International Conference on Knowledge Management and Information Sharing
340

machine defined on the knowledge set
K
and
)(PL
is the language generated by the product, in
other words, a sequence of elements in
K
, noted as:

⎪
⎭
⎪
⎬
⎫
⎪
⎩
⎪
⎨
⎧
==
∈∧∈
=
njni
QEKkkkk
PL
P
j
P
i
P
n
PP
,..,2,1,0;,..,2,1
);()(|...
)(
21
where
T
j
E
is the state in the evolution process of the
product
P .
We have
PT ⇔
, that is to say, each general
task
T
links to a product development process of
product
P
. For example, the task of “design a
chair”, means to realize a product design process,
from the initial state “an idea” to the final state
“details of the chair”.
3.3 The Application Process of the
OFCK Model
Thanks to the OFCK model, some issues in
knowledge management can be simplified to the
problem of seeking solutions from a knowledge base
when the general object and constraints are given
(shown in Figure 2). For example, in a case of
designing a chair, the OFCK model can be applied
to seek the knowledge required.
1. Decomposition. The general task of
“designing a chair” can be decomposed into
sub-tasks such as considering the architecture,
considering the cost, considering the recycle
possibilities, considering aesthetics, etc.
2. Mapping. Once the atomic tasks are clarified
in the design process, the atomic knowledge
can be found by the use of the FCK map.
3. Retrieving. The required knowledge can be
retrieved through the ILK or imported from
the extern.
4 CONCLUSIONS AND
PERSPECTIVES
This paper has proposed the OFCK model to
describe the relationship of objects, constraints and
knowledge, which helps to find the appropriate
knowledge to a given general object in the context of
certain constraints. The main idea of the paper is to
find the appropriate knowledge to a given task.
Further studies may aims at the optimization
procedure of the product design process, in other
words, how to accomplish the tasks in a more
efficient way. Another research opportunity is about
the improvement of the characterization of the
product design process. When some processes have
recursions, the task decomposition process will be
much more complicated and the mapping process
should be reconsidered.
REFERENCES
Bernard, A., Tichkiewitch, S., 2008. Methods and tools for
effective knowledge life-cycle-management. Springer:
Berlin.
Doumeingts, G., Ducq, Y., Vallespir, B., Kleinhans, S.,
2000. Production management and enterprise
modeling. Computers in Industry 42(2-3): 245-263.
Gravemeier, V., Wall, W. A., 2008. A space–time
formulation and improved spatial reconstruction for
the “divide-and-conquer” multiscale method.
Computer Methods in Applied Mechanics and
Engineering 197(6-8): 678-692.
Labrousse, M., 2004. Proposition of a unified conceptual
model for the dunamic enterprise knowledge
management. PhD. thesis, Ecole Centrale de Nantes,
France .
Lester, N., Zobel, J., Williams, H., 2006. Efficient online
index maintenance for contiguous inverted lists.
Information Processing & Management 42(4): 916-
933.
Merlo, C., Girard, P., 2004. Information system modelling
for engineering design co-ordination. Computers in
Industry 55(3): 317-334.
Nagalingam, S. V., Lin, G. C. I., 2008. CIM—still the
solution for manufacturing industry. Robotics and
Computer-Integrated Manufacturing 24(3): 332-344.
INTEGRATION OF KNOWLEDGE MANAGEMENT IN PRODUCT DESIGN PROCESS
341