Download 44ševtsenko

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
5th International DAAAM Baltic Conference
"INDUSTRIAL ENGINEERING – ADDING INNOVATION CAPACITY OF LABOUR
FORCE AND ENTREPRENEURS"
20–22 April 2006, Tallinn, Estonia
DEVELOPMENT OF IDSS FOR COLLABORATIVE NETWORK OF
PRODUCTION ENTERPRISES
Ševtšenko, E., Karaulova, T., Kyttner, R.
Abstract: The scope of this research work
is to describe the approach for the
Intelligent Decision Support System (IDSS)
design and analysis. The intelligent
decision support system for the network of
collaborative enterprises will provide
decision support for the participants and
enable reliable improvement in data
transfer. The system will enable to achieve
new vision of collaborative work between
enterprises and increase the profitability
level of a participant. The paper includes
overview of IDSS analysis. It includes the
basics of DSS and describes the main areas
of usage.
1. IDSS BASICS
In today environment of international
competition it is extremely important to
make things better, faster, cheaper than
your competitors.
Data flow
IDSS of production companies
IDSS of subcontractor
companies
IDSS of customer
companies
Data flow
Data flow
IDSS of logistic
companies
Data flow
IDSS of vendor
companies
Collaborative
network of
enterprises
Data flow
reveals that this difference in system
efficiency is not due to advantages in
manufacturing equipment. Company can
still manage to maintain a stronger position
throw better management practices and a
more systemic approach than its
competitors.
2.
THEORETICAL
PROBLEM SOLVING
BASE
FOR
IDSS are a specific class of computerized
information system that supports business
and
organizational
decision-making
activities.
As the main model for the decision support
the strategic (aggregate) planning (AP) is
proposed [1]. Aim is to maximize the profit
for the whole network of enterprises. The
model was approved by computation
example. It was confirmed that model can
be used for resource constrains analysis
and
long-range
production
plan
optimisation in the collaborative network
of enterprises.
It has been clear to most organizations for
some time now that the intelligent
exploitation of data at all levels within the
company is the key differentiator between
success and failure. [2].
Fig. 1. IDSS in the collaborative network
of enterprises
The management decisions today are made
in the conditions of fast changes.
The aim of IDSS is to support effectiveness
of the collaborative network of enterprises
(Fig.1), which enables to achieve max
profit and min lead-time. Examination of
the production practices of the companies
That is why the data required for
management solutions is dynamically
collected from External and Internal
sources to Data Warehouse. IDSS is able to
queue for required data and to transform
this
data
into
solution
(Fig.2)
Data analysis & Knowledge discovery
Management
agent
Data
Warehouse
Supply
agent
Internal
Data
Manufact.
agent
Data mining
External
Data
Solutions
Planning
OLAP
Planning
agent
Queries
Data source
Processes
management
IDSS
Optimization
…
…
Fig. 2. IDSS role in the production management process
IDSS will support the work of the agile
enterprises throw quick response to the
changes in the work environment.
In figure 3 are given steps for knowledge
discovery for decisions making.
Data Presentation
Data Mining
Data Exploration
Data Warehouses
Data Sources
Data Mining and Knowledge
Discovery
Making Decisions
Increasing
potential to
support decisions
Fig. 3. Knowledge discovery for decisions
making
Recognizing that needless data is found
within most enterprises, organizations have
implemented some form of “data
warehousing” to bring this data together
and allow efficient enterprise reporting and
analysis. Key information needs to reach
multiple decision makers within the
organization.
Inconsistencies
and
inefficiencies result when information is
not distributed and shared across the
organization.
3.
ANALYSIS
PROBLEM
OF
EXISTING
The next question is how to design and
implement such IDSS. It is believed that
one factor related to unsuccessful
implementation is that the functions and
underlying mechanism of the relevant
manufacturing systems in question have
not been sufficiently analysed and
understood by the people concerned in
manufacturing industry [3]. It is required to
define the source of the existing problem.
This is the key to the problem solving.
Different people, depending on their
involvement, will have different views on
what the system in question does. When
assembling a system, therefore, the analyst
should first of all be very careful about the
viewpoint he or she takes.
System AS-IS model is the initial point for
the reveal the bottlenecks of the system
work. In TO-BE model we try to dispose
this limitations for improvement of the
system (Fig.4).
AS-IS
model
Additional
functional
requirements
TO-BE
model
Fig. 4. Carriage of analysis
During
analysis
process
analysis
transforms the AS-IS model into TO-BE
model. The result will be model of future
system, which will be accepted by key
users of future system. The process model
is the basis for business process simulation,
rationalization, and reengineering. [6].
4. IDSS MAIN USES
Decisions are made and business
intelligence is garnered only with the
combination of the output of the decision
support tools, human judgment and
intuition, and the ability to put the
information spit out by tools into a context
of information that is much wider than any
data warehouse, transaction processing
system, knowledge repository can handle.
conceptual model will show how future
system should look like, and what tasks it
should be able to perform. The states of
future system are changing continuously
over time, so that model would be
dynamic. The system is intelligent, or the
system will act in order to fulfil the
required tasks. As open system it will have
the ability to adapt to its environment by
altering the structure and processes of its
internal environment. The external
environment of the IDSS system includes
data warehouse of collaborative enterprises
and even channel to IDSS of partner
enterprise can be established after the
cooperation agreement is achieved. The
future model will be discrete, because
system variables change in a stepwise
fashion. In other words system will
perform tasks accordingly to the real
situation at the moment. So the behaviour
of the system can be analysed by using a
methodology
called
discrete-event
simulation. System also could be classified
as stochastic, because it is characterized by
random properties [3].
IDSS main uses are [4]:
 To convey information in a more
digestible manner;
 To compare information about
customers,
products,
cost/profit
centres, financial accounts;
 To compare the same type of
information in different time periods;
 To check performance versus formal
and informal goals or constraints;
 To grab a little piece of information
out of a large volume of information;
 To confirm and sometimes to discover
trends and relationships;
 To help advocate a position;
 To provide data for what if analysis or
a forecast.
5. IDSS MODEL DEVELOPMENT
It is proposed to start with IDSS system
model development, which will be based
on collected information. Further analysis
will be done through testing and
improvement processes of existing model.
5.2. Input-output analysis
Classification of inputs outputs is important task to start with.
5.1. Defining the system type
To start with it is required to determine the
category of future system.
Due to its no physical existence the future
system model will be conceptual. The
C1
Market price
Enterprise
equipment resources
Due date
Agregate planning theory
Forcast of sales
I1
I2
I3
I4
Resource data
Product data
General ledger
Providing of
information
for strategic
optimization
Master production
shedule
O1
Bills-of materials
Analysis of
constrains
Material
requirements
planning
O2
Resource plan
I5
Sales and
purchase data
A1
Required item
Strategic
optimisation
of activities
Item cost
Enterprise data DB
Quontity of items
Optimised
long-range plan
Max profit
I6
ERP
planning
O4
Reports
A3
A2
Processing
information
routing and machines
required
M2
ERP network
M1
IDSS
Fig. 5. IDEF0 diagram of production network management process
O3
Product cost
variability
We should make clear what are the inputs
and outputs of our system. It is proposed to
use IDEF methodology for input-output
analysis.
Models are useful for understanding how
the system works, and how parameters
move. Especially it is needful during the
system development or for collaboration
and communication different parts of
system.
The IDSS accommodates data from several
business functions. The diagram of
production network management process is
proposed. The process is performed by
ERP and IDSS. Systems have different
tasks to perform and are able to exchange
information with each other. Description
IDSS business functions is introduced in
figure 5 by using IDEF0 method.


At the beginning the IDSS could be
thought as some kind of ´black box`. We
know nothing about inside processes and
have a look at external environment throw
inputs and outputs. We know what
information we have, and what the results
must be. Then we can go deeper and start
with analysis of extrinsic processes.
Feedback analysis of the IDSS could be
performed in the way of analysis how
system outputs are related to the system
inputs.
5.4. Communication
Principle of communication is one of the
prerequisites for successful control.
Various pieces of information are needed
to enable decision to be made. The output
of the decision-making function will again
be information in the form of instructions
or constrains. These must be channelled to
the intended destinations to initiate logical
control actions. A big amount of data is
required for strategic decisions making. In
case if some information is lost or gone to
wrong address the result could be wrong
decision made and money lost as the
outcome.
collaborative network of enterprises.
5.3. Agent interactions and feedback
control
The concept of an intelligent agent is used
as the essential technique for analysis and
modelling. Distributed object technology
allows computing systems to be integrated
so that objects or components work
together across a machine and the network
boundaries.
The behaviour of the agents of the IDSS
system can be divided:
 The agent will try to solve the problem
independently;
Models of Agents
Planning
Agent
IDEF, UML
Collaborative network of
enterprises
EXCEL files
Results of simulation
Decision support
The agent will communicate with other
intrinsic agent;
The agent will communicate with other
extrinsic agent.
Management
Agent
Supply
Agent
Manufacturing Prod.development
Agent
Agent
Enterprise 1
Enterprise 2
Enterprise 3
Enterprise 1
data
Enterprise 2
data
Enterprise 3
data
IDSS
Fig. 6. Decision making scheme by using agent approach
5.5. Description and use of a prototype
system model
It is proposed to use system prototype in
order to understand the real system
involved in a particular problematic
situation. As an example IDSS production
network management prototype is used to
test different agent relationships. It is
convenient to analyse the communication
process between agents during the whole
decision support process (Fig.7).
6. COMPUTER SIMULATION
THE MAIN IDSS PROCESSES
Simulation is a very active branch of
computer science, which consists of
analysing the properties of theoretical
models of the surrounding world. As a rule
for the system analysis the system
simulation is used. On the basis of the
simulation results decision will be done
related to the needed changes in
configuration at the system or performance
at its elements.
The process of subcontractor selection was
tested in the Arena software. The result is
graphical representation of subcontractor
selection process and process reports.
7. IMPLEMENTATION OF IDSS
It is proposed to divide the implementation
process into several stages: diagnostics,
analysis and design, implementation of
IDSS system, IDSS solution support
(Fig. 8).
Constrains analysis
agent
Searching of
subcontractor
agent
Collaborative
network strategic
prod.plan
optimisation agent
IDSS
mediator agent
Strategic
optimization of
activities
ERP agent of
own enterprise
Manufacturing
export agent
Supply export
agent
ERP agent of
subcontractor
enterprise
Supply import
agent
Planning import
agent
Import/Export agents
Providing of
information for
strategic optimization
* project skope
definition;
* specification of
customer
believes and
requirements
* solution analysis
* IDSS
requirements
analysis
* Design of IDSS
Decision
How to
continue?)
Single enterprise
strategic prod.plan
optimisation agent
Analysis and
Design
Decision
(should we
continue?)
Diagnostics
Manufacturing
import agent
OF
Implementation
of
IDSS system
* development,
* testing,
* training
* IDSS go life
IDSS solution
support
*help-desk
*analysis of
requirements
*implementaion of
new possibilities
Decision
(where to
continue?)
The multi layer of the decision making
scheme was developed (figure 6.) Each
agent has a proper structure, which may be
described by using several standards, for
example the IDEF (Integrated Definition
language) or UML (Unified Modelling
Language). Different agents are used to
perform defined tasks. Agents are
independent and able to collect required
information from the network of
collaborative enterprises. Optimisation
processes are performed by MS Excel
solver and in is the base for decision
support.
ERP agent of
own
enterprise
ERP agents
ERP
planning
PROCESSES
Fig. 7. The prototype of communication
between agents
Prototype system model tells us what a
properly designed and implemented system
should look like. In association with the
prototype
system
model,
system
identification and description is one of the
essential skills, which any system analyst
must learn.
Fig. 8. Steps of IDSS implementation
7.1. Diagnostics
The diagnostics is the process of defining
the project scope. It must ensure that IDSS
is compatible with the software used in the
enterprise. The scheme of application used
in the enterprise must be made. Then it
will be decided what applications the IDSS
must be able to communicate with. The
total approximate cost and duration of
IDSS implementation will be estimated.
The result will be decision to start the
project.
7.2. Analysis and design
During
this
stage
the
enterprise
requirements and problems will be
analysed. It will be decided what role will
be given to IDSS system. It will be
estimated what are the critical processes of
the company and what will be the base for
decision support. The result will be the
prototype of the desired system, which
includes all the requirements IDSS must
satisfy.
7.3. Implementation of IDSS system
During this stage the IDSS system will be
tested, improved and implemented. All key
users will be trained to use new system. It
is possible to make some minor changes in
the IDSS design.
7.4. IDSS solution support
At this stage the working system is
supported. The improvement process is
active, and when the scope of new
developments required increased the new
implementation project will be started.
8. CONCLUSION
4. IDSS development and implementation
stages:

Diagnostics;

Analysis and design;

Implementation;

Solution support.
Acknowledgements Hereby we would like
to thank the Estonian Science Foundation,
for the grant 6795 enabling us to carry out
this work.
8. REFERENCES
1. Wallace J.Hopp; Mark L.Spearman.
Factory
physics:
foundation
of
manufacturing
management,
Irwin/McGraw-Hill, 2001.
2. Hummingbird,
Decision
Support
Solution, Hummingbird Ltd., 2004.
3. B. Wu, Manufacturing systems design
and analysis, Chapmann & Hall, ISBN
0 412 58140 X, 1994.
4. Greenfield L, What Decision Support
This paper introduces the vision of possible
way for IDSS development (Figure 9).
1.Analysis of the
problem
IDSS
4. IDSS development develop
and implementation
ment
stages
stages
2.Development of
the IDSS model
3. Simulation of the
main processes
Figure 9. IDSS development stages
It is proposed that development of IDSS
system could be based on predefined tasks:
1. Analysis of the problem that IDSS
must be able to solve.
2. Development of the IDSS model:
 Input output analysis;
 Agent definition;
 Agent communication;
 Using of prototype.
3. Computer simulation of main processes
Tools are Used For, LGI Systems
Incorporated, 2005
5. Cques Ferber, Multi-agent systems,
Addison-Wesley, 1999.
6. Salvendy G., Handbook of Industrial
Engineering:
Technology
and
Operation Management, JOHN WILEY
 SONS, INC., New York, ISBN 0471-33057-4, 2001.
9. CORRESPONDING AUTHOR
M.Sc. Eduard Ševtšenko
Department of Machinery, TUT, Ehitajate
tee 5, 19086, Tallinn, Estonia
Phone: +372 620 3265
E-mail: [email protected]