1990 Fiscal Year Final Research Report Summary
Development of Expert System of Supporting Bridge Design with Learning Ability
| Project/Area Number |
01850103
|
|---|
| Research Category |
Grant-in-Aid for Developmental Scientific Research (B).
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
土木構造
|
|---|
| Research Institution | Kyoto University |
|---|
Principal Investigator |
FURUTA Hitoshi Kyoto Univ., Dept. of Civil Engin., Assoc. Prof., 工学部, 助教授 (70109031)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
SAWARAGI Tetsuo Kyoto Univ., Dept. of Precision Engin., Instructor, 工学部, 助手 (10187304)
SUGIURA Kunitomo Kyoto Univ., Dept. of Civil Engin., Instructor, 工学部, 助手 (70216307)
SUGITO Masata Kyoto Univ., Dept. of Transportation Engin., Instructor, 工学部, 助手 (60115863)
AKIYAMA Takamasa Kyoto Univ., Dept. of Transportation Engin., Lecturer, 工学部, 講師 (70159341)
|
|---|
| Project Period (FY) |
1989 – 1990
|
| Keywords | Expert system / Learning / Bridge / Aesthetic factor / Production Rule / Design Supporting system |
|---|
| Research Abstract |
In the design of bridges, it is necessary to consider various design factors such as safety, economy, beauty and maintenance. To obtain an adequate bridge design which satisfies the above all requirements, intuition and experience may play an important role. However, it is very difficult to accumulate such experience enough to make a good design from the point of practical design situation. In this study, an attempt is made to develop an expert system which can support for inexperienced engineers to design a bridge. Especially, attention is paid to the stage of design planning, in which aesthetic factors are important as well as safety and economical factors. In the evaluation of the aesthetic factors, it is inevitable to include the qualitative evaluation which sometimes varies due to the subjectivity of designers. The present system has such a learning ability that it can automatically generate rules necessary for the inference. By the automatic generation of rules, it is possible to reduce the loads necessary to collect the sufficient amount of rules. This procedure is also useful in generating the rules from the source of qualitative information.
|
