Transcript PowerPoint プレゼンテーション

CSM2002, July 15-17,2002, IIASA, Austria
Study on the Metasynthesis of Data, Information,
Model and Expert Opinions
Gu Jifa
School of Knowledge Science
Japan Advanced Institute of Science
and Technology
Tang Xijin
Institute of Systems Science
AMSS
Chinese Academy of Sciences
I. INTRODUCTION
For some complex system problems it is often the case we couldn't
only use the data and models to solve them, We have to combine
the human judgment (qualitative) and mathematical models
(quantitative), even more we have to use the Metasynthesis from
qualitative to quantitative approach proposed by Prof. Qian et al.
This Metasynthesis approach suggested that when solve the
problems arisen in the open giant complex system, we must
integrate the Data, Information, Knowledge, Models and expert
opinions with the help of advanced computer technology.
In order to realize this approach there were a lot of
Chinese researchers engaged in it. In June of 1999
The National Natural Science Foundation of China
(NSFC) had approved a main program titled in
"Metasynethetic systems with combination between
man and machine for decision support of
macroeconomics" (1999-2003).
14 universities and research institutes
around 50 professors, research staffs and
graduated students
II. ORGANIZATION OF MAIN PROGRAM
There are four subprojects in this program:
P1. Information and model systems for
macroeconomics and their functions;
P2. Metasynthetic systems with combination between
man and machine and supporting environment;
P3. Metasynthetic method systems and systematology
researches for decision support of macroeconomics;
P4. Knowledge discovery system (KDD) and cognitive
researches for macroeconomics.
The main task of subproject P3 is to find the way to
realize this metasynthesis approach.
The conceptual model of this project is shown as below.
Data
Model
Method
Experience
Macro
economic
s
Decision support for
macroeconomics
After two years research the four research organizations
in this subproject-P3 Shanghai Jiaotong University,
Xian Jiaotong University , Beijing Normal University
and Institute of Systems Science had found some
techniques and ways to realize the task individually.
Shanghai Jiaotong University had developed the Multi-agent
technology to integrate the forecasting models for
macroeconomics and Feasible Desirability Method (FDM) to
combine the expert opinions,
Xian Jiaotong University had developed the design of
"common brain" to support unstructured group process.
Institute of Systems Science provided the Group
Argumentation Environment (GAE) and Metasynthesis
Reconstructability Analysis (MRA). to make consensus building
After several discussions we had integrated all these techniques
and methods altogether and used the web technology to
connect them.
III.Design of metasynthesis approach
3.1 General design
Synchronic stage I (Meeting I)
Asynchronous Stage (Analysis)
Synchronic stage II (meeting II)
3.1 Meeting I (Rough Discussion)
3.1.1. Electronic Common Brain
3.12. Brainstorming
3.1.3. KJ Method
3.1.4. Delphi Method
3.1.5. Group Argumentation Environment (GAE)
3.1.1. Electronic Common Brain
Zhang Pengzhu and his colleagues from Xian Jiaotong University are
now developing a prototype of the Electronic Common Brain, or simply
Common Brain, which are useful for the group decision making process.
When the experts attend the discussion they may have two kind of
discussion Labs: online Discussion Lab and offline discussion Lab. In the
first stage people may classify, retrieve, aggregate rough information,
then let the information be systematic and input into the database attached
to offline discussion Lab. Then in the offline discussion Lab people may
make statistics, analyze, verify and reach primitive consensus. Finally
people again go to online discussion Lab to run the deeper discussion.
They also investigate the problem how to support the group decision
making process with unstructured information and propose a tree-structure
for information organization in unstructured group process.
3.1.5. Group Argumentation Environment (GAE)
Tang and Liu in Institute of Systems Science develop a prototype
environment for Group argumentation called as Group Argumentation
Environment (GAE). It includes three modules: brainstorming
argumentation room (BAR), automatic affinity list and nominal group
(NG) . A synchronous meeting may be taken in BAR. In this stage the
different and divergent thinking are encouraged The automatic affinity
may help people to analyze and visualize qualitative ideas expressed by
experts in a two dimensional map. NG may be used in the second
synchronous stage for converging the opinions from experts
3.2 Analysis
3.2.1 Model I
Econometric Model; Time Series Model
3.2.2 Model II
Multi-agent Model
3.2.3. Model III
New economic model
3.2.4. Model from outside
3.2.5. Model Integration
3.3. Meeting II
3.3.1. Metasynthesis Reconstructibility Analysis(MRA)
3.3.2. Feasible Desirable method (FDM)
3.3.3. Consensus Building
3.3.1. Metasynthesis Reconstructibility Analysis(MRA)
•Klir proposed system reconstructability analysis in 1976. In
China we had developed this methods both in theory and
practice. We combine this analysis with metasynthesis, and
wish apply it in solving the complex system problems, like the
comprehensive evaluation of social-economical-environmentalecological impacts of the three gorges project, biotechnology,
engineering design, forecasting and risk analysis etc. recently
we are planning to use it in the macroeconomic decision
making. After modification we can use the data, information
and knowledge altogether.
•Shu has developed this method as Metasynthesis
Reconstructability Analysis and used them to forecast the
growth rate of GDP in China.
Year (1)
(2)
(3)
1987 11.6
11.0057
11.385
1988 11.3
12.515
11.8706
1989 4.1
12.3064
4.5313
1990 3.8
6.53981
3.82484
1991 9.2
13.5869
8.73321
1992 14.2
14.2655
14.326
1993 13.5
14.6255
12.075
1994 12.6
12.7926
13.206
1995 10.5
7.35298
9.9316
1996 9.6
7.06571
9.09417
1997 8.8
7.2239
8.03538
1998 7.8
7.08204
8.07829
(1) GDP(growth rate) in Yearbook
(2) GDP forecasted
(3) GDP forecasted with considering the knowledge
Working flowchart of Metasynthesis Approach
Data
Information
Knowledge
Case
Consensus II
Meeting I
Consensus I
Analysis
FDM
Model I
Model II
Models from outside
Meeting II
Model III
Model
Integration
MRA
Working flowchart of Metasynthesis Approach (with organizations)
(ISS)
Data(710,TU)
Information(AM)
Knowledge
Case(HU)
(XJU)
Meeting I
(ISS)
Consensus I
Consensus II
Analysis
(710,AM)
(SJU)
FDM
Model I
(PU)
Model II
(BNU,SJU)
Models from outside
Meeting II
Model III
(SJU,ISS)
Model
Integration
(ISS)
MRA
3.3.3. Consensus Building
3.3.3.1 Definition
3.3.3.2 Useful methods and tools for building consensus Method
3.3.3.3 Effective and efficiency meeting
3.3.3.4.DMTMC system
3.3.3.1. Definition of consensus
1. Agreement of opinions
2. Conformity of different parts in a system
# Consensus in various mathematical meanings
Physics
MCDM (Pareto)
Fuzzy
Mathematical Statistics
Game theory
Risk
Rough
Consensus under fuzziness( Kacpryzyk, Nurmi and Fedrizzi edited
1997, Kluwer Academic Publishers)
Fuzzy preferences
Fuzzy majorities
Degree of consensus(CON): most of the relevant individuals agree
as to almost all of the important alternatives
Fuzzy aggregation: OWA(Ordered weighted averaging operator);
Fuzzy linguistic quantifiers
Degree of Q1/Q2/I/B-consensus
Degree of /Q1/Q2/I/B-consensus
Degree of s/Q1/Q2/I/B-consensus
I :Important; Q1:number of pairs of relevant alternatives;
Q2:number of pairs of important individuals; :degree sufficient
s:degree of strength
Additional agreement indicators: Contribution to consensus( CTC);
Personal consensus degree( PCD); Detailed personal consensus
Degree( DPCD); Contribution to consensus for options( OCD);
Option consensus degree( OCD)
Mediator (Polish Systems Research Institute)
Borda definition
Condorcet definition
Coombs definition
Copeland definition
French Election type definition
Hare definition
Minmax definition
Nanson definition
Plurality definition
3.3.3.2 Useful methods and tools for building consensus
Email
Internet
Intranet
Teamware
Groupware
Cooperative system
Coordination system
Collaborative system
BBS ( Bulletin Board System)
Teleconferencing system
Computerized conferencing
EMS (Electronic Mail System)
Electronic conference system
GDSS ( Group Decision Support System)
CSCW (Computer Supported Cooperative Work)
Decision room
Discussion Hall
Virtual Reality
3.3.3.3. Effective and efficiency meeting
• Meeting type
• Schedule for meeting
• Facilitation
• Mediation
3.3.3.4. DMTMC system
Data
Data
Meeting
Meeting
Tool
Tool
Consensus
Consensus
Method
Method
IV. Conclusion
Though the mentioned key project is under progress, but we found some
specific features during running the project. They are:
1) Follow some system methodologies, like metasynthesis system approach,
Wuli-Shili-Renli (WSR) system approach and Spiral Propulsion system
approach;
2) Provide economic data base, economic information, data mining;
3) Design the meeting, using Common Brain ,GAE etc ;
4) Provide different economic models, and model integration;
5) Provide metasynthesis methods, which may synthesize the data, information,
models and experience, e.g. NGT, AHP, System reconstruction, and some
consensus methods, such as voting, group decision methods;
6) Study on the some theoretical topics related to the economic complex systems.
Reference
1.
Tang X.J. and Liu Y.J., A prototype environment for Group Argumentation, presentation on MCS2002, August 7,2002,
Shanghai
2.
Cheng Shaochuan, Sun JingIe, Liu Ming-de, On support paradigm of group decision argument, J. of Systems Engineering,
16(5) 366-370, 2002
3.
Hu Daiping, Wang Huanchen, Building forecasting model system in workshop for hall of metasynthetic engineering to
support macroeconomy decision, J. of Systems Engineering, 16(5) 335-339, 2002
4.
Shu Guangfu, Meta-synthetic system reconstruction and application in macro-economic researches, J. of Systems Engineering,
16(5) 349-353, 2002
5.
Tang Xijin, Model integration, J. of Systems Engineering, 16(5) 322-329, 2002
6.
Shen Huizhang, Wang Huanchen, Methodology based evolution modeling for macroeconomics systems analysis, J. of Systems
Engineering, 16(5) 389-393, 2002
7.
Ge Xinyuan, Waang Dahui, Yuan Qiang, Fang Fukang, General multi-sector dynamic economic model andits reasonability
analysis, J. of Systems Engineering, 16(5) 397-401, 2002
8.
Gu Jifa , On synthesizing opinions-how can we reach consensus, J. of Systems Engineering, 16(5) 366-370, 2002