| Project/Area Number |
11480070
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B).
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | HOKKAIDO UNIVERSITY |
|---|
Principal Investigator |
SATOH Ken Hokkaido University., Grad.School of Eng., Asso.Prof., 大学院・工学研究科, 助教授 (00271635)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
IWANUMA Koji Yamanashi University, Fac.of Eng., Asso.Prof., 工学部, 助教授 (30176557)
SAKAMA Chiaki Wakayama University, Fac.of Sys. Eng., Asso.Prof., システム工学部, 助教授 (20273873)
INOUE Katsumi Kobe University, Fac.of Eng., Asso.Prof., 工学部, 助教授 (10252321)
|
|---|
| Project Period (FY) |
1999 – 2000
|
| Project Status |
Completed (Fiscal Year 2000)
|
| Budget Amount *help |
¥11,100,000 (Direct Cost: ¥11,100,000)
Fiscal Year 2000: ¥3,900,000 (Direct Cost: ¥3,900,000)
Fiscal Year 1999: ¥7,200,000 (Direct Cost: ¥7,200,000)
|
|---|
| Keywords | multi agent system / incomplete information / speculative computation / abduction / abductive logic programming / default reasoning / 仮説論理型プログラミンク / 仮設論理型プログラミング / 仮設推論 |
|---|
| Research Abstract |
We presented speculative computation by abduction in multi-agent systems and showed that this method can be used for distributed problem solving under incomplete communication environments.
The idea of the method is as follows :
- When communication is delayed or failed, then we use a default hypothesis as a tentative answer and continue computation.
- When some response is obtained, we check consistency of the response and the current computation. If the response is consistent, then we continue the current computation ; else if the response is inconsistent, we seek an alternative computation.
We proposed a proof procedure of speculative computation in a master-slave multi-agent systems and showed that it returns correct answers for the perfect model semantics in a stratified logic program. Based on the proof procedure, we made a simiulator on parallel machines to show effectiveness of this method.
|
