1996 Fiscal Year Final Research Report Summary
Identification of the old-grown broad-leaved forest dynamics with non-linear engineering method
| Project/Area Number |
07806021
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
林学
|
|---|
| Research Institution | SHIZUOKA UNIVERSITY |
|---|
Principal Investigator |
NOGAMI Keiitirou SHIZUOKA UNIVERSITY,AGRICULTURE,ASSOCIATE PROFESSOR, 農学部, 助教授 (50150511)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
OOSAKA Okihiro SHIZUOKA UNIVERSITY,AGRICULTURE,ASSOCIATE PROFESSOR, 農学部, 助手 (20252166)
YUASA Yasuo SHIZUOKA UNIVERSITY,AGRICULTURE,ASSOCIATE PROFESSOR, 農学部, 助手 (70022069)
|
|---|
| Project Period (FY) |
1995 – 1996
|
| Keywords | Non-linear engineering method / Fractal / Genetic algorithms / Fuzzy theory / Neural network / Multiple forest land use planning / Kansei / Artifical intelligence |
|---|
| Research Abstract |
Fractal geometry, genetic algorithms, neural network (parallel distributed processing) and fuzzy theory, as the typical representative concepts of non-linear engineering method, were introduced in this research. It has been shown that these techniques are effective measures in comprehension and modeling complicatednature objects and social phenomenon. Meanwhile, they have been scarcely applied in forestry.
The fractal characteristics of stand structure (trees location, species diversity, diameter distirbution) of old-grown broadleaved were examined and verified. The results suggested that 3 factors all follow fractal distribution.
The relationship of fractal characteristics of tree crowns structure and maximun density curves was indicated. The result showed that fractal dimension of tree crowns is closely associated with the ecological constant of maximum density curve.
The power of neural network was demonstrated in applying it to the problem of multiple forest land use and visual quality ascertain. In order to bulid the sturucture of neural network, an unusual method was introduced. Namely, based on the genetic algorithms, the number of input, the intermediate layr and the neuron number which formed intermediate layr, the type of threshold function were optimized.
For the purpose that to promote this initial research, problems on the collection of necessary essential data were discussed.
Application of non-linear engineering method were prospected. The necessity for developing the growth process prediction model of broad-leaved forests and the fuzzy expert system which based on a serial researches was suggested.
|
