This article was downloaded by: [Politechnika Warszawska]
On: 18 October 2014, At: 10:58
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office:
Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

International Journal of Computer
Integrated Manufacturing

Publication details, including instructions for authors and subscription
information:

http://www.tandfonline.com/loi/tcim20

Information design for SAIL 1

J. Zhang
Published online: 08 Nov 2010.

To cite this article: J. Zhang (1996) Information design for SAIL 1, International Journal of Computer
Integrated Manufacturing, 9:1, 22-31, DOI:

10.1080/095119296131788

To link to this article:

http://dx.doi.org/10.1080/095119296131788

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis, our agents,
and our licensors make no representations or warranties whatsoever as to the accuracy,
completeness, or suitability for any purpose of the Content. Any opinions and views expressed
in this publication are the opinions and views of the authors, and are not the views of or
endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall
not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages,
and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection
with, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial
or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or
distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use
can be found at

http://www.tandfonline.com/page/terms-and-conditions

Inform ation design for SAIL 1

J. ZHANG, K. B. CHUAH, H. M. CHEUNG and Z. DENG

Abstract.

SAIL 1 at City University of Hong Kong is to be

developed into an integrated m anufactu ring system with
¯ exible m anufacturing and m aterial handling capabilities.
To m eet the life-cycle requirem ents of SAIL 1 from prelimin-
ary design through the system implem entation and operation,
CIM -OSA m ethodology is being used to de® ne its top-down
integration strategy and formulate th e bottom -up develop-
m ent tactics. Based on brief analyses about the system con® g-
uration and information ¯ ow structure of SAIL 1, this paper
presents a data m odelling strategy to m odel its information
system , which organizes th e inform ation resources from the
system physical con® guration, system operation control and
system operation m onitoring, respectively. By using a CIM -
O SA approach to abstract, classify and de® ne the object views
and enterprise objects for the information view of th e SAIL 1
system at the requirem ent de® nition level, an object-oriented
semantic data m odel was constructed in accordance with the
system requirem ents and data m odelling strategy, from which
th e conceptual schem a for th e future information system of
SAIL 1 can be derived.

1. Introduction

As a kind of production and management philosophy

developed over a decade or so, computer integrated
m anufacturing (CIM) has been widely accepted by both
industry and academia as the new generation manufac-
turing strategy. It is recognized as a m ost promising
m ethod to bring m anufacturing enterprises into the
next century, and a powerful weapon that is able to
improve product quality and enhance com petitive
power in the global economic m arket. To keep pace
with and further the advances of CIM technologies so as

to promote and facilitate their application in the local
m anufacturing industry, the Departm ent of M anufactur-
ing Engineering at City University of Hong Kong
(City U) is developing two substantial CIM facilities
called System -level Application and Integration Labora-
tories (SAILs). SAIL 1 is to be developed into an
integrated system with ¯ exible m anufacturing and
m aterial handling capabilities, while SAIL 2 is a ¯ exible
assembly system (FAS). The research and development
goal for SAILs is to provide an integrated m anufacturing
environm ent for student learning, staff teaching and
research, and local industrial services in CIM (Cheung

et al

., 1992). As one of the principal R&D institutions in

the CIM area of China’s high-tech R&D Program me, the
Department of Manufacturing Engineering at Nanjing
U niversity of Science and Technology (NUST) has con-
tributed to the research and development of SAILs at
City U .

The SAIL 1 system at present consists of m ore than

10 pieces of com puter num erical control (CNC) equip-
m ent with a wide range of degrees of automation,
m achining functions and m aterial handling techniques.
To integrate these currently stand-alone equipments
into an organic system com patible with operation in a
CIM environment, a CIM Open System Architecture
(CIM-OSA) m ethodology, proposed by the CEC/
ESPRIT program me, is being applied to de® ne SAIL
1’s top-down system integration strategy and formulate
its bottom-up development tactics. The very ® rst phase
of such a methodical approach consists of careful and
thorou gh system requirem ent analysis and system m od-
elling in function, information and resource. This paper
deals with the inform ation modelling for the SAIL 1
system.

It has been increasingly believed that inform ation

integration is the ® rst and also the m ost important
problem to be solved during the system integration
procedure because the information ¯ ow, or data ¯ ow,
is the unique m edium which ties together all the opera-
tional functions and activities that are com pleted by
different com ponents of the system, including hardware
and software. A well-designed and properly implemen-
ted inform ation system is one that can provide multiple

0951-192X/ 96 $12.00

€

1996 Taylor & Francis Ltd

INT. J. CO M PU TER INTEGRATED M ANU FAC TU RIN G,

1996,

VO L.

9,

N O .

1, 22 ± 31

Authors:

J.

Zhang:

Associate

Professor,

Department

of M anufacturing

Engineering, Nanjing University of Science and Technology, Nanjing 210014,
P.R China. Currently a post-doctoral visiting scholar at Dept. of M echanical,
Aerospace and Nuclear Engineering, University of California, Los Angeles, Los
Angeles, CA 90024-1597, USA . K. B. Chuah: Lecturer, Department of
M anufacturing Engineering, City University of Hong Kong, HK. H. M.
Cheung: Principal Lecturer, Department of M anufacturing Engineering, City
University of Hong Kong, HK. Z. Deng: Professor, Department of M anufacturing
Engineering, Nanjing University of Science and Technology, Nanjing 210014,
P.R China. Currently a visiting professor at Narvik Institute of Technology,
Norway.

Downloaded by [Politechnika Warszawska] at 10:58 18 October 2014

inform ation services throughout the m anufacturing
system at different levels to all system components
which need it in the m ost effective and ef® cient
m anner possible (Zhang and Liu 1989). Effective here
m eans that the information provided has to be com -
plete, accurate and just in time, while ef® cient means
that the m anner by which information is provid ed is
resource conservative and cost-effective.

To m eet the aforementioned requirements for an

inform ation system of an integrated m anufacturing
system , it is very important to have an appropriate
m ethodology to analyse, organize and construct the
system -wide inform ation m odel from which the infor-
m ation system could be derived or mapped out. In this
paper the inform ation analysis of SAIL 1 will ® rst be
brie¯ y conducted, based on which, the information
m odelling strategy will be investigated and presented
accordingly. Then the CIM -O SA m ethodology will be
used to m odel the inform ation view of the SAIL 1 system
and an object-oriented semantic data m odel for SAIL 1
will be presented.

2. Inform ation analysis for SAIL 1

It is known from the R&D plan and the function

design for SAIL 1 that the prospective SAIL 1 system is
com posed of three FMS (¯ exible m anufacturing
system )-type m anufacturing cells, two DNC (distributed
num erical control)-type m anufacturing cells and a
m aterial handling system (Fung and Cheung 1992).
The difference between an FM S-type cell and a DNC-
type cell is that there exist both inform ation ¯ ow and
m aterial ¯ ow in the FM S-type cell but only information
¯ ow in the DNC-type cell. From the viewpoint of infor-
m ation integration, the inform ation system for these two
types of manufacturing cell can be m odelled in almost
the sam e way. The inform ation model of a DNC-type cell
can be derived from an information m odel of an FM S-
type cell by rem oving material ¯ ow related objects from
it.

As com pared with the hierarchical control structure

of a CIM system proposed by NBS (now NIST) of the
U SA, it is found from the system requirement and
structure analyses that the prospective SAIL 1 is a typical
shop¯ oor m anufacturing system. As shown in Figure 1,
SAIL 1 corresponds to the lower four levels in the NBS-
CIM S structure. In order to have a thorough under-
standing of the characteristics of information ¯ ow in
SAIL 1 and to construct a feasible inform ation m odel for
it, it is very important to collect all inform ation require-
m ents in all respects at each level of the SAIL 1 system .
Based on the inform ation requirem ents analysis, func-
tion design for SAIL 1 and by reference to the hierarch-
ical control structure of NBS-CIMS, the inform ation ¯ ow
in SAIL 1 is abstracted as shown in Figure 2. Further
analysis of data input and output for different levels in

the system was made in more detail by Chuah

et al.

(1993).

3. Data modelling strategy for SAIL 1

The objective to m odel all possible inform ation con-

tents for SAIL 1 is to develop a consistent, shareable,
complete and non-redundant inform ation system which
should be able to serve all users and applications both
inside and outside the system. To m ake the constructed
data m odel m ore compatible with a CIM environm ent,
and m ore conform able to system requirem ents, it is
very important to determine an appropriate data m od-
elling strategy because this makes a notable impact on
the effectiveness and reliability of the m odelled
inform ation system.

The inform ation ¯ ow of SAIL 1, as shown in Figure 2,

is organized hierarchically and analysed conceptually in
accordance with the hierarchical control structure of

Information design for SAIL 1

23

Figure 1. NBS-CIM S structure and SAIL 1 structure.

Figure 2. Information ¯ ow structure in SAIL 1.

Downloaded by [Politechnika Warszawska] at 10:58 18 October 2014

NBS-CIMS. The information m odel and the information
system for SAIL 1 can also be constructed hierarchically
to conform to its functional structure. In this way,
the data input and output m echanisms and data rela-
tionships have to be established and de® ned separately
at different levels. However, it has to be pointed out
that a great deal of information in the m anufactu ring
system , e.g. design data ® les of parts and NC programs
for part m achining and material handling, are charac-
terized by transmissibility throughout the whole system .
These data with transmissibility are supposed to be
shared by many different functions at different levels
in the system. The degree of sharing is directly pro-
portional to the length of data transm itting path
(Zhang and Liu 1989). To m eet this practical require-
m ent during the system operation process, it is very
important to guarantee data consistency and complete-
ness, and at the sam e time, shareability of these data
with transmissibility at different locations. It can be
naturally inferred that the hierarchically constructed
inform ation m odel for SAIL 1 m ay result in data
redundancy, data inconsistency and data incom plete-
ness in its further developed inform ation system
because of the separation of data linkages am ong
different levels.

Considered from the viewpoint of the system con-

® guration, SAIL 1 is com posed of two `item s’, nam ely
hardware and software. Hardware implies the physical
con® guration

of

the

system ,

which

consists

of

tangible objects such as equipm ent and com puters.
Software, on the other hand, implies those intangible
system com ponents that are responsible for imple-
m enting system operation controlling and m onitoring
functions and activities. The com bination of system
hardware and software conceptually form s different
system levels of the hierarchical control structure as
shown in Figure 1. As the ® rst step in organizing the
inform ation

resources effectively

and

completely,

the inform ation system of SAIL 1 is modelled from the
following three entry points as shown in Figure 3.

d

The physical con® guration of the SAIL 1 system ,
which

is

related

to

the

physical

resource

m anagement and scheduling issues of both the
SAIL 1 and even the City U .

d

System operation control, which is related to the
m anufactu ring planning and scheduling issues.

d

System operation m onitoring, which is responsible
for assuring the effective and reliable system
operation.

3.1.

Physical con® gu ration oriented m odel

To implement the system operation control and

m onitoring strategy, a data model is needed to provide
all relevant system applications and users with the
required inform ation abou t the system’s physical
con® guration. This m odel can be organized and
constructed based on the following three categories.

(1)

Equ ipm ent category.

Com prises all objects that can

operate or m ove under the control of NC program s and
carry out speci® c operational functions, e.g. part
m achining or m aterial handling, during the m anu-
facturing process. In the SAIL 1 system, this m odel is
composed of machine tools, coordinate m easuring
m achine (CM M), autom ated guided vehicle (AGV),
robo t, washing m achine, drying m achine, com puters,
part loading/unloading station, and tool loading/
unloading station. This categorized m odel will be
able to provide relevant users and applications with
basic information such as equipment identi® ers,
nam e and type, as well as more detailed inform ation
such as the operational parameters of the speci® c
equipment.

(2)

R esource category.

Involves those physical objects

that are unable to provide program med or autom ated
services to system users or application, but m eanwhile
are de® nitely needed for the system operation. Objects
grou ped into this category are cutting tools, ® xtures,
pallets, buffers, m aterial, and personnel who are
required passively at different working positions. Inform-
ation such as resource identi® er, nam e and their
detailed param eters are very important to the m anufac-
turing process planning and scheduling departm ents.

(3)

System

category.

To

identify

properly

the

inform ation ¯ ow in the SAIL 1 system and strengthen
the system operation control and managem ent, the
components that constitute the hierarchical structure
of the SAIL 1 system have to be grouped into this
category. This m odel will comprise the logically existed
components such as shop¯ oor, cell and workstation, and
physically existing components such as central tool base
and warehouse. The m odelling of these system-related
components will facilitate the planning, scheduling,
control and m onitoring functions and activities inside
and outside the SAIL 1 system .

J. Zhang

et al.

24

Figure 3. Data m odelling strategy for SAIL 1.

Downloaded by [Politechnika Warszawska] at 10:58 18 October 2014

3.2.

Operation control oriented m odel

According to the operational requirement for SAIL 1

(Cheung

et al.

1992), it should be able not only to be

integrated and operate in a CIM environment, but also
to operate independently. So it is necessary for SAIL 1 to
be able to accept production ord ers, shop¯ oor produc-
tion plans and engineering data either from the factory
level or directly from system users by m eans of the m an ±
m achine interfaces as shown in Figure 2.

The data m odel should contain all the objects that are

necessary for controlling and scheduling the system
operation, which involves production orders, shop¯ oor
production plans, cell production plans, m anufactu ring
process plans, operations, NC-programs, job groups,
jobs, schedules, order sheets, resource requirem ents
and supplying plans that are related to such objects as
personnel, cutting tools, ® xtures and material.

3.3.

Operation m onitoring oriented model

In

an

advanced

manufacturing

system,

m any

operation-related decision-making activities such as
planning, scheduling, com municating and ord er dis-
patching are carried out according to the accurate and
timely data reports provided by the monitoring system
com posed of a great number of sensors and data
processors, throughout the system. Such data capturing
and reporting are to guarantee the operational reliabil-
ity of every com ponent of the system . Objects to be
m odelled into this data model involve the following
three categories of report.

(1)

Statistic reports:

usually the resource utilization

statistics, provide the upper levels with decision-
m aking references for the allocation of production
tasks and the generation of production plans for the
lower levels. Statistical information m ay be generated at
the shop¯ oor, cell and workstation levels in SAIL 1.

(2)

Capacity reports:

usually consist of capacity and

capability data are also the reference inform ation for
the upper levels to generate or decom pose production
plans for the lower levels. Capacity reports will be
generated from cell, workstation and equipment levels.

(3)

Status reports:

are the m ost important information

to guarantee the reliability of system operation and will
re¯ ect the current status of such objects as equipm ent,
workstation, cell, part and other resource objects. The
status reports will provide the required current status
data to all relevant components and activities to effect
their real-time control.

As a particular object served by a num ber of activities

and other objects in the SAIL 1 system, a part may
appear in all of these aforem entioned data m odels

because there exist m any interrelationships between
the part and other objects in these m odels.

After the determ ination of the data modelling

strategy for the information system of SAIL 1, it is very
important to have an appropriate data modelling
m ethodology and tool to organize the inform ation
and implement the data m odel. To m eet the R&D
requirem ents for SAIL 1, it is strongly recomm ended
that the inform ation system of SAIL 1 be modelled into
one with open system architecture possessing openness
in both time and space dom ains to m eet the speci® c
requirem ents of SAIL 1 such as its compatibility with a
CIM environm ent, operational independence and ¯ exi-
bility, and system expandability and recon® gurability.

4. CIM -OSA methodology and inform ation view for
SAIL 1

4.1.

CIM -O SA and information view

Since inform ation is the lifeblood of a CIM system and

the binding glue which ties together the various system
functions and components, a great deal of attention has
been paid to its m odelling and m anagement (Su 1989,
Smith and Wang 1989, Vernadat

et al.

1989). Although

several tools and m ethodologies have been proposed to
guide the development of inform ation system, they
usually cannot satisfy all the requirements of the
inform ation system that supports manufacturing envir-
onm ents (Chadha

et al

. 1991). They lack the com pre-

hensive form al fram ework necessary to guide the
m odelling process through to the implementation of
an inform ation system that supports the m anufacturing
system integration. In addition, m any of these existing
m ethodologies were developed for business inform ation
and hence need to be enhanced to suit m anufacturing
m odelling requirem ents.

CIM-OSA, proposed by the AM ICE consortium of the

CEC/ESPRIT program m e, is a system life-cycle oriented
CIM

architecture

which

de® nes

an

integrated

m ethodology to support all phases of a CIM system,
from

requirem ents

speci® cation,

through

system

design, implem entation, operation and maintenance,
and even system m igration towards CIM-OSA solution
( Jorysz and Vernadat 1990a). Based on the different
generic extents, i.e. generic, partial and particular, the
CIM -OSA m ethodology conducts system requirem ents
de® nition, design speci® cation and implementation
description from such respects as function, inform ation,
resource and organization, which are described as the
function view (FV), inform ation view (IV), resource view
(RV) and organization view (OV). Among these views,
the FV is the base to generate and analyse the other
three, while the IV plays a pivotal role in the integration
and com munication of all these views (Zhang and Deng

Information design for SAIL 1

25

Downloaded by [Politechnika Warszawska] at 10:58 18 October 2014

1992). CIM-OSA architecture consists of a modelling
fram ework that is a three-dimensional cube of 36 m od-
elling blocks corresponding to different m odelling tasks
at different stages of the system life-cycle, an integrated
infrastructure, and a set of enterprise design software
packages used to support a CIM system requirement
de® nition, design and implementation. The com po-
nents and their interrelationships in the CIM-OSA inte-
grated environm ent are as shown in Figure 4.

The object± entity relationship attribute (OERA)

approach has been adopted by CIM -OSA m ethodology
as the basis for the inform ation view m odelling para-
digm. It makes use of an object-oriented m odel for
high level sem antic modelling at the requirem ent de® -
nition level. At the design speci® cation level, a m ore
precise, data-oriented semantic m odel is used for infor-
m ation system design, which is based on the entity
relationship attribute (ERA) approach. Finally, at the
implementation description level, CIM-OSA methodol-
ogy provides translation m echanisms to transform an
ERA m odel into relational, hierarchical or network
database structure ( Jorysz and Vernadat 1990b).

The inform ation m odelling paradigm used for CIM -

OSA information view m odelling is com pliant with the
three-schema approach de® ned by ANSI. The basic
m odelling constructs used at the various level of CIM -
OSA information view and the transforming mechan-
isms between different levels are sum m arized as shown
in Figure 5 ( Jorysz and Vernadat 1990b). The infor-
m ation design for the SAIL 1 system at the system
analysis and preliminary design phase corresponds to

inform ation m odelling at the requirem ents de® nition
m odelling level in the CIM-OSA inform ation view as
shown in the Figure 5.

4.2.

Information m odelling for SAIL 1

The goals of inform ation m odelling for SAIL 1 at the

requirem ent de® nition level are:

d

to de® ne enterprise data and inform ation require-
m ents using a structurally object-oriented m odel;

d

to capture suf® cient inform ation detail to facilitate
the derivation of these requirem ents into a consis-
tent inform ation m odel of the enterprise inform a-
tion system.

The inform ation language proposed to business users

by CIM-OSA m ethodology is essentially based on the
concepts of enterprise object, object view and inform a-
tion element. In order to capture m ore of the semantics
of information, concepts of object abstraction mechan-
isms and integrity rules were added into the inform ation
m odelling language set ( Jorysz and Vernadat 1990b).
The usual modelling procedure is started from the
determination of object views, then the abstraction
and classi® cation of enterprise objects, ® nally the

J. Zhang

et al.

26

Figure 4. Integrated environm ent for CIM -OSA.

Figure 5. CIM-O SA information view m odelling constructs.

Downloaded by [Politechnika Warszawska] at 10:58 18 October 2014

object-oriented sem antic data m odels are constructed
based on the object abstraction m echanisms. The infor-
m ation elements are determined concurrently at the
® rst two stages, and the integrity rules are em bodied in
the constraints in the object views. Before starting to
m odel the inform ation view of the SAIL 1 system,
however, it is necessary to present brie¯ y the concept
of enterprise activity in the CIM-OSA function view
because it is the only hub from which we can map the
inform ation requirem ents in the function view into the
inform ation view.

4.2.1.

Enterprise activity.

As is known from the descrip-

tion of the CIM -O SA function view, the enterprise
activity is a construct that is used to de® ne the atom ic
functionality in a domain of the enterprise. Its diagram -
m atic representation is given in Figure 6, from which the
inform ation requirements for the functional operation
of the system can be abstracted. An enterprise activity
m ay have the following three kinds of input and output:

·

Function inputs/outputs.

The function inputs are

object views of the enterprise objects which are
operated on during the execution of the enterprise
activity. The function outputs are also object views of
the enterprise objects which are the results of the
enterprise activity.

·

Control inputs/outputs.

The control inputs are

object or information elements which control and/or
constrain the execution of the enterprise activity.
The control output consists of a set of information
elements that de® ne the ending status of the enterprise
activity.

·

Resource inputs/outputs.

The resource inputs are

enterprise objects which are utilized by the enterprise
activity to perform its transform ation function. The
resource outputs describe the resource objects as they
are returned to the resource pool after execution of the
enterprise activity.

It is known from the m apping m echanisms between

the function view and inform ation view at the CIM -OSA
requirem ent de® nition level (Zhang and Deng 1992)
that only function inputs, function outputs and control
inputs can be abstracted and de® ned as the object views
in the information view. The control outputs are usually
status inform ation which is used to control the business

process in the CIM-OSA function view. The resource
inputs are not considered at the requirement de® nition
level, while the resource outputs are returned status
inform ation required by the CIM-OSA resource view.
The object views, therefore, are one of the hooks linking
the CIM -O SA function and inform ation views at the
requirem ent de® nition level.

Information m odelling at the requirement modelling

level for the SAIL 1 system is com posed of the following
two entry points:

d

extract the object views from those identi® ed inputs
and outputs of enterprise activities in the m odelled
function view;

d

determ ine the enterprise objects by both grouping
the extracted object views and foreseeing non-
structured inform ation com ing from the dom ain
description, the resource view and even organiza-
tion view.

4.2.2.

Object views in SAIL 1.

An object view is the descrip-

tion of a particular aspect of an enterprise object. It
therefore describes a speci® c set of properties and opera-
tions owned by an enterprise object from a given stand-
point. The concept of object view will facilitate further
evolution of the inform ation system , since global descrip-
tion uses only enterprise objects, while users and appli-
cations use only object views, which matches well the
three-schema m ethodology.

By de® nition, an object view is a restricted projection of

an enterprise object. The restriction can be som e infor-
m ation elements or the enterprise object itself. An object
view is a recursive construct de® ned by:

d

unique identi® er;

d

nam e;

d

the enterprise object it focuses on, called leading
object;

d

other enterprise object to get properties from ,
called related objects;

d

other related object views;

d

list of properties consisting of inform ation elem ents
and/or other object views.

The template of object view is shown in Figure 7 ( Jorysz
and Vernadat 1990b). Based on the function design and
the aforementioned inform ation m odelling strategy for
SAIL 1, 100 object views are de® ned in the set of object
view for SAIL 1. Nam es of these de® ned object views are
listed in Appendix A.

4.2.3.

Enterprise objects in SA IL 1.

The concept of enter-

prise object is the result of a unifying trend to m odel the
whole of industrial reality through a few universal con-
cepts at a high level of abstraction. Enterprise objects are

Information design for SAIL 1

27

Figure 6. Enterprise activity.

Downloaded by [Politechnika Warszawska] at 10:58 18 October 2014

constructs which describe generalized entities of the
enterprise system at the user level (i.e. things that can
be conceptualized or thought about as being a whole),
and are characterized by descriptive properties and by
operations which can be applied to them .

The enterprise object allows the description of highly

structured inform ation such as products, parts and
schedule, as well as the bulk information such as reports,
m easures, etc. Enterprise objects are m ade up of lower
level enterprise objects (simply called sub-objects) and/
or of inform ation elements. Objects and sub-objects are
linked together by m eans of object abstraction mechan-
isms described later. An enterprise object is constructed
by its unique identity, its name and descriptive proper-
ties. It can be recursively built by using sub-objects or
object views as properties. The template of an enterprise
object is shown in Figure 8.

By m eans of a generalized concept, 41 enterprise

objects for the SAIL 1 system are extracted and classi® ed

from the previously de® ned object views. It is noticed
that m any of the object views have to be m odelled
directly as enterprise objects because of their original
characteristics as a whole. Names of these de® ned
enterprise objects are listed in Appendix B.

4.2.4.

Semantic data m odel for SAIL 1.

The ultimate aim

of data modelling for SAIL 1 is to construct a data m odel
from which an information system can be physically
derived to serve all the system operations. The CIM -
OSA m ethodology adopts an object-oriented modelling
approach to model the information contents at the
requirem ent de® nition level because it is one in which
the data are grouped into packages that occur naturally,
which m akes the data m odel compatible with the way
people think about the real world. It allows for data
abstraction,

inheritance,

inform ation

hiding

and

dynamic binding. Here object abstraction m echanisms
are used to connect enterprise objects to one another to
form a semantic information m odel at the requirement
de® nition level. In order to m ake the m odelled data

J. Zhang

et al.

28

OBJECT VIEW TEM PLATE

Type:

[relevant category - select from list]

Identi® er:

[OV-unique num ber]

N am e:

[m nem onic nam e]

D esign Authority:

[nam e of person with authority to
design/ m aintain this particular
instan ce]

N ature:

[Ph ysical or Informational]

D escription:

[short description]

Leading Object:

[nam e of m ain Enterprise O bject to
which the Object View refers to]

Related O bject:

[optional list of nam es of Enterprise
O bjects directly related to the Object
View]

Referenced Object: [optional list of nam es of Enterprise

O bjects which m ay be referred to in the
O bject View de® nition]

Referenced Views:

[optional list of nam es of O bject Views
which m ay be referred to in the O bject
View de® nition]

Properties:

[list of Properties of the O bject View
de® ned either as:

an Information Elem ent nam e

an Object View nam e

<

identifier

>

``:Setof ’’ followed by either an Inform ation

Elem ent nam es or an Object View nam e.

<

identifier

>

is

th e property nam e]

Constraints:

[optional descriptive list of constraints
applicable to th e O bject View]

Figure 7. Object view tem plate.

EN TERPRISE O BJECT TE M PLATE

Type:

[relevant category - select from list]

Identi® er:

[EO-unique num ber]

N am e:

[Nam e of the Enterprise O bject]

D escription:

[sh ort description]

D esign Authority:

[nam e of person with auth ority to
design/m aintain th is particular
instan ce]

Abstraction Relationships:

[list of sem antic object relationships in
th e form:

Isa:

[list of nam es of Enterprise Objects
which are a G eneralization of this
particular instance.
The nam e of the Enterprise Object
being de® ned cannot be used]

Partof:

[list of nam es of Enterprise Objects in
which th is particular instanc e is involved
in an Aggregation. The nam e of the
Enterprise Object being de® ned cannot
be used]

Properties:

[list of properties of th e Enterprise
O bject, each one being either:

an Information Elem ent nam e

an Enterprise O bject nam e

<

identifier

>

``:Setof ’’ followed by either a list of Infor-

m ation Elem ent nam es.

<

identifier

>

is th e property

nam e]

Figure 8. Enterprise object tem plate.

Downloaded by [Politechnika Warszawska] at 10:58 18 October 2014

m odel unam biguous in m eaning and non-redundant in
structure, two kinds of object abstraction m echanisms,
generalization and aggregation, are used to construct
the sem antic data model for the SAIL 1.

Generalization is an abstraction mechanism which

enables an individual object type to be thought of as a
m ore generic object typ e. This abstraction mechanism is
used to organize enterprise objects into a hierarchy
through a partial order relationship, often referred to
as the IS-A hierarchy (Brachm an 1985).

Aggregation refers to an abstraction mechanism in

which an enterprise object type is regarded as a con-
junctive collection of sub-component objects, which
de® nes one-to-many or m any-to-m any associations
between enterprise object types. This is known as the
PARTOF link between enterprise objects and enterprise
object classes.

Relationships between enterprise objects are also

adopted in the semantic data m odel for the SAIL 1
system to de® ne the so-called LINKEDTO relationships.
They can be of typ e one-to-one, one-to-many or m any-to-
m any.

Based on the enterprise objects de® ned, object

decomposition and data m odelling strategy stated in
Section 3, an object-oriented sem antic data m odel is
constructed as shown in the Figure 9. This data m odel

uses the generalization and aggregation abstraction
m echanisms, and relationships to describe sem antic
links among enterprise objects.

5. Conclusions

This paper deals with the information system modelling

issues of SAIL 1, an integrated system with ¯ exible
manufacturing and material handling capabilities. Based
on a brief analysis of the system con® guration and
information ¯ ow structure of the SAIL 1, it presents a
data modelling strategy to organize the information
resources in the SAIL 1 from three respects:

d

system physical con® guration that is related to its
resource m anagement and scheduling issues;

d

system operation control which is related to the
m anufactu ring planning and scheduling;

d

system operation m onitoring which is responsible
for assuring the reliability and effectiveness of the
system operation.

Since the data m odelling strategy organizes the infor-

m ation resource based on the characteristics of the

Information design for SAIL 1

29

Figure 9. O bject-oriented data m odel for SAIL 1.

Downloaded by [Politechnika Warszawska] at 10:58 18 October 2014

inform ation contents rather than strictly following the
physical control structure of the system, its implementa-
tion in future information systems will de® nitely
improve data consistency and data completeness, and
reduce data redundancy.

CIM-OSA is a system life-cycle oriented CIM system

architecture which de® nes an integrated m ethodology
to support all phases of a CIM system from requirement
speci® cation, through system design, implem entation,
operation and m aintenance. Because of the character-
istics possessed by its architecture, and its theoretical
and standardization background, the application of
CIM -OSA m ethodology in the system design and m od-
elling for SAIL 1 will assure the m odelled system meets
system requirem ents such as compatibility in a CIM
environm ent, operational independence and ¯ exibility,
and system expandibility and recon® gurability, while the
application of CIM-OSA methodology in information
m odelling will de® nitely improve the quality of the
m odelled inform ation system in its openness, effective-
ness and reliability, consistency and com pleteness.

Based on the m apping m echanisms between the

CIM -OSA function view and information view, 100
object views were extracted, which re¯ ect the system
requirem ents for information, and 41 enterprise
objects were classi® ed and de® ned, which represent
the system con® guration of SAIL 1. Finally, an object-
oriented semantic data m odel was constructed in accor-
dance with the system requirements and the proposed
data modelling strategy, from which the conceptual
schem a for the future inform ation system of SAIL 1
can be correspondingly derived.

References

B

R AC HM AN

, R. J., 1985, W hat is IS-A is and Isn-t: An analysis of

taxonom ic links in sem antic networks. IEEE Computer, 16, 30±
36.

C

H A DHA

, B., J

A ZBU T IS

, G., W

A NG

, C. Y., and F

ULTO N

, R. E., 1991,

An appraisal of m odeling tools and m eth odologies for
integrated m anufacturing information system s. Proceedings
of the 1991 ASM E International Computer in Engineering Con-
ference, Santa C lara, CA.

C

H E UN G

, H. M ., F

UN G

, Y. K., and L

I

, S. H ., 1992, CIM in

H ong Kong Ð

its education, research, developm ent, and

industrial application. Proceedings of International Sym -
posium on Computer Integrated Manufacturing Systems, Beijing,
China.

C

H UA H

, K. B., F

U NG

, R., C

H E UN G

, E., Z

HA N G

, J., and D

EN G

, Z.,

1993, Information design for SAIL 1 Ð

Part 1: information

design. Proceedings of the 9th International Conference on
Computer Aided Production Engineering. N anjing, C hina.

F

UN G

, Y. K., and C

H EU NG

, H . M ., 1992, The current state of

com puter integration on m anufactu ring m anagem ent
and control. Proceedings of International Sym posium on
Computer Integrated M anufacturing Systems. Beijing, China.

J

O R YSZ

, H . R., and V

E RN A DAT

, F. B., 1990a, CIM -OSA Part

1: total enterprise m odeling and functional view. Interna-
tional Journal of Computer Integrated M anufacturing, 3, 144±
156.

J

O R YSZ

, H . R., and V

ERN A DAT

, F. B., 1990b, CIM -O SA Part 2:

information view. International Journal of Computer Integrated
M anufacturing, 3, 157 ± 167.

S

M ITH

, G. W., and W

AN G

, M ., 1987, M odeling CIM system s Part

III: an intelligent database design environm ent for the
design and implem entatio n of a CIM information system .
Computer Integrated M anufacturing Systems, 2, 99 ± 107.

S

U

, S. Y. W., 1986, M odeling integrated m anufacturing data

with SAM *. IEEE Computer, 19, 34 ± 49.

V

ERN A DAT

, F., D

I

L

E VA

, A., and P

IERC ARL O

, Q., 1989, O rganiza-

tion and information system design of m anufactu ring envir-
onm ents: th e new

M * approach. Computer Integrated

M anufacturing Systems, 2, 69 ± 81.

Z

H A NG

, J., and D

E N G

, Z., 1992, Study on the m apping m echan-

isms from function view to information view in CIM -OSA.
Proceedings of CIM S-China ’92, Shenzhen, China.

Z

H A NG

, J., and L

IU

, Y. E., 1989, The analysis of data

structure and th e application of database in CIM S. Proceed-
ings of the 2nd International Conference on Computer Aided
Drafting, Design and M anufacturing Technology, Hangzhou,
China.

J. Zhang

et al.

30

1. Production Order
2. Production Order

Identi® er

3. Shop Production Plan
4. Shop Production Plan

Identi® er

5. Process Plan
6. Process Plan Identi® er
7. Operation
8. Operation Identi® er
9. NC-Program

10. NC-Program Identi® er
11. Task Allocation Plan
12. Task Allocation Plan

Identi® er

13. Cell Production

Plan

14. Cell Production Plan

Identi® er

15. Job Group
16. Job Group Identi® er
17. Job
18. Job Identi® er
19. Schedule
20. Schedule Identi® er
21. Order Sheet
22. Order Sheet

Identi® er

23. Resource Requirem ent

Plan

24. Resource Requirement

Plan Identi® er

25. Resource Supplying

Plan

26. Resource Supplying

Plan Identi® er

27. Report
28. Report Identi® er
29. Statistic Report
30. Statistic Report

Identi® er

31. Status Report
32. Status Report

Identi® er

33. Capacity Report

34. Capacity Report

Identi® er

35. Shop¯ oor Identi® er
36. Shop¯ oor Status
37. Cell Identi® er
38. Cell Status
39. Warehouse Identi® er
40. Warehouse Status
41. Central Tool Base

Identi® er

42. Central Tool Base

Status

43. Workstation

Identi® er

44. Workstation Status

Appendix A. Defined object views of SAIL 1

Downloaded by [Politechnika Warszawska] at 10:58 18 October 2014

Information design for SAIL 1

31

45. Equipm ent Identi® er
46. Equipm ent Status
47. Machine Tool
48. Machine Tool

Identi® er

49. Machine Tool Status
50. AGV
51. AGV Identi® er
52. AGV Status
53. Robot
54. Robot Identi® er
55. Robot Status
56. Washing Machine
57. Washing Machine

Identi® er

58. Washing Machine

Status

59. Drying M achine
60. Drying M achine

Identi® er

61. Drying M achine

Status

62. CMM
63. CMM Identi® er
64. CMM Status
65. Com puter
66. Com puter Identi® er
67. Com puter Status
68. Part L/UL Station
69. Part L/UL Station

Identi® er

70. Part L/UL Station

Status

71. Tool L/U L Station

72. Tool L/UL Station

Identi® er

73. Tool L/UL Station

Statu s

74. Part
75. Part Identi® er

76. Part Status
77. Workpiece
78. Workpiece

Identi® er

79. Workpiece Status
80. Resource
81. Resource Identi® er
82. Resource Status
83. Personnel
84. Personnel Identi® er
85. Personnel Status

86. M aterial
87. M aterial

Identi® er

88. M aterial Status
89. Tool
90. Tool Identi® er
91. Tool Status
92. Fixture
93. Fixture Identi® er
94. Fixture Status
95. Pallet
96. Pallet Identi® er
97. Pallet Status
98. Buffer
99. Buffer Identi® er

100. Buffer Status

Appendix B. De® ned enterprise objects of SAIL 1

1. Shop¯ oor
2. Cell
3. Warehouse
4. Central Tool Base
5. Workstation
6. Equipment
7. M achine Tool
8. Autom ated Guided Vehicle
9. Robot

10. Washing Machine
11. Drying M achine
12. Coordinate M easuring M achine
13. Com puter
14. Part L/U L Station
15. Tool L/U L Station
16. Resource
17. Personnel
18. Material
19. Part
20. Workpiece
21. Tool

22. Fixture
23. Pallet
24. Buffer
25. Process Plan
26. Operation
27. NC-Program
28. Production Order
29. Shop Production Plan
30. Task Allocation Plan
31. Cell Production Plan
32. Job Group
33. Job
34. Schedule
35. Order Sheet
36. Resource Requirement Plan
37. Resource Supply Plan
38. Report
39. Statistic Resport
40. Status Report
41. Capacity Report

Downloaded by [Politechnika Warszawska] at 10:58 18 October 2014