| Project/Area Number |
25330261
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Kyushu University |
|---|
Principal Investigator |
Hatano Kohei 九州大学, 附属図書館, 准教授 (60404026)
|
|---|
| Co-Investigator(Renkei-kenkyūsha) |
TAKIMOTO Eiji 九州大学, システム情報科学研究院, 教授 (50236395)
KIJIMA Shuji 九州大学, システム情報科学研究院, 准教授 (70452307)
NAGANO Kiyohito 公立はこだて未来大学, システム情報科学部, 准教授 (20515176)
|
|---|
| Project Period (FY) |
2013-04-01 – 2016-03-31
|
| Project Status |
Completed (Fiscal Year 2015)
|
| Budget Amount *help |
¥4,810,000 (Direct Cost: ¥3,700,000、Indirect Cost: ¥1,110,000)
Fiscal Year 2015: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2014: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2013: ¥1,950,000 (Direct Cost: ¥1,500,000、Indirect Cost: ¥450,000)
|
|---|
| Keywords | オンライン予測 / 離散構造 / 離散最適化 / スケジューリング / 順序 / Bradley-Terry モデル / ブースティング / バンディット問題 / 機械学習 / ランキング / オンライン密度推定 / 順列 / ランキング予測 / 近似アルゴリズム |
|---|
| Outline of Final Research Achievements |
Online prediction problems of combinatorial objects arise in many applications such as scheduling, ranking and network optimization and so on. For individual problems, there are several results on robust and efficient online prediction methods. However, there is no unified approach on how to design efficient algorithms and there seems rooms for improvements.
In this project, we consider a generic approach that converts offline algorithms for combinatorial optimization into prediction algorithms. An advantage of our approach is that the prediction performance reflects the approximation performance of offline algorithms. Therefore, good offline algorithms imply good online prediction algorithms.
|
