| Project/Area Number |
10680389
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | UNIVERSITY OF THE RYUKYUS |
|---|
Principal Investigator |
ONAGA Kenji UNIVERSITY OF THE RYUKYUS, FAC. OF ENG., PROFESSOR, 工学部, 教授 (90029869)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
NAKAMURA Morikazu UNIVERSITY OF THE RYUKYUS, FAC. OF ENG., ASSOCIATE PROFESSOR, 工学部, 助教授 (80237437)
|
|---|
| Project Period (FY) |
1998 – 1999
|
| Project Status |
Completed (Fiscal Year 1999)
|
| Budget Amount *help |
¥3,300,000 (Direct Cost: ¥3,300,000)
Fiscal Year 1999: ¥800,000 (Direct Cost: ¥800,000)
Fiscal Year 1998: ¥2,500,000 (Direct Cost: ¥2,500,000)
|
|---|
| Keywords | DISTRIBUTED ALGORITHM / STABLE MARRIAGE / MATHING RECONSTRUCTION / DISTRIBUTED STABLE MARRIAGE PROBLEM / AUTONOMOUS DISTRIBUTED SCHEDULING / PARALLEL MACHINE PROBLEM / FLEXIBILITY / MOBILE ROBOTS / 二並行マッチング再構成 / 自律分散アルゴリズム / マッチングリコンストラクション / 木構造神経回路網 |
|---|
| Research Abstract |
In this research, we prove that in distributed stable marriage problems two concurrent partner exchanging can be performed. Moreover, for parallel machine problems, we develop an autonomous distributed scheduling scheme based on message passing.
In the stable marriage problem, for one problem instance (preference list) there exist generally more than one stable matching. This fact is also true in case of distributed version of this problem. We use the notation "matching reconstruction" for traversing autonomously from one stable matching to another one. In our previous works, we treated the case that only one matching reconstruction process can exist at the same time. However, this condition is very strict for cooperation works of multiple mobile robots since they have to wait until the previous matching reconstruction is completed and some mutual exclusion mechanism is needed. In this research, as a better solution for autonomous distributed environments, we prove that two concurrent matching reconstruction can be performed successfully.
For the parallel machine problems, we develop an autonomous distributed scheduling scheme. Computer simulation shows that proposed scheme can reduce more the total delay from the deadline of the given products than traditional approaches. Moreover a mixed version of our scheme and the traditional one can improve the solution quality. Flexibility for additional jobs and machine failure is also observed in the experiment, that is, rescheduling is quickly performed by using information obtained during the previous schedule computation. The quality of approximated solutions is also reasonable even if it is worse than that of genetic algorithms. However, the computation time is quite small comparing to genetic algorithms. Our proposed scheduling scheme can be applicable to cooperation of mobile robots.
|
