| Project/Area Number |
25280087
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Partial Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Waseda University |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
栗原 聡 電気通信大学, その他の研究科, 教授 (30397658)
廣津 登志夫 法政大学, 情報科学部, 教授 (10378268)
|
|---|
| Co-Investigator(Renkei-kenkyūsha) |
FUKUDA Kensuke 国立情報学研究所, アーキテクチャ科学研究系, 准教授 (90435503)
YAMAMOTO Hitoshi 立正大学, 経営学部, 准教授 (70328574)
|
|---|
| Project Period (FY) |
2013-04-01 – 2017-03-31
|
| Project Status |
Completed (Fiscal Year 2016)
|
| Budget Amount *help |
¥11,180,000 (Direct Cost: ¥8,600,000、Indirect Cost: ¥2,580,000)
Fiscal Year 2016: ¥2,990,000 (Direct Cost: ¥2,300,000、Indirect Cost: ¥690,000)
Fiscal Year 2015: ¥2,990,000 (Direct Cost: ¥2,300,000、Indirect Cost: ¥690,000)
Fiscal Year 2014: ¥3,120,000 (Direct Cost: ¥2,400,000、Indirect Cost: ¥720,000)
Fiscal Year 2013: ¥2,080,000 (Direct Cost: ¥1,600,000、Indirect Cost: ¥480,000)
|
|---|
| Keywords | 分散協調問題 / 負荷分散 / 提携形成問題 / チーム編成問題 / ノルム / 大規模分散システム / マルチエージェントシステム / 強化学習 / 提携形成 / 分散システム / 自己組織化 / 知的エージェント / 分散資源割り当て / マルチエージェント機械学習 / 組織ネットワーク / リソース割当 / 進化ゲーム / 分散協調資源割当問題 / マルチエージェント学習 / 組織化 / 自律ネットワーク / 最後通牒ゲーム / 協調ゲーム / 互恵性 / 分散資源割当問題 / 組織 |
|---|
| Outline of Final Research Achievements |
We proposed a distributed task allocation method in which a large number of agents with different capabilities construct a group through processing in a bottom-up manner and allocate appropriate resources / tasks within their groups to maximize the efficiency and the use of their capability. We also evaluated the proposed method and investigated the effect of the interaction between agents on the entire performance. We introduced the different agent’s behavioral strategies and by selecting one of them by local learning, we tried to analyze the efficiency by using an abstract repetitive game that was also proposed in this study. We found that (1) by learning behavioral strategies, agents could autonomously build a robust organizational structure consisting a number of groups based on their reciprocity, and (2) the mixed structure of reciprocal agents as well as rational agents that interact with various members in different groups based on the expected income increases efficiency.
|
