2013 Fiscal Year Final Research Report
Constructing Search Algorithms Taking Account of Global Multimodality
| Project/Area Number |
23500273
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Sensitivity informatics/Soft computing
|
|---|
| Research Institution | Tokyo Institute of Technology |
|---|
Principal Investigator |
ONO Isao 東京工業大学, 総合理工学研究科(研究院), 准教授 (00304551)
|
|---|
| Project Period (FY) |
2011 – 2013
|
| Keywords | 大域的最適化 / 進化計算 / 大域的多峰性 / 大谷構造 / 関数最適化 / 実数値進化計算 |
|---|
| Research Abstract |
In this study, we proposed new search algorithms to efficiently find good solutions in globally multimodal space. It is required that, in the globally multimodal space, search algorithms efficiently find promising big valleys and efficiently search the best solution in each big valley. In order to achieve this requirement, we proposed a framework for finding big valleys and several real-coded evolutionary algorithms for searching in a big valley. The proposed framework iteratively executes real-coded evolutionary algorithms and efficiently find new big valleys by using a history of search regions. We confirmed that the proposed framework with a real-coded genetic algorithm called AREX/JGG succeeded in finding the optima that the conventional methods failed to find on globally multimodal benchmark functions. We also confirmed that the proposed real-coded evolutionary algorithms for searching in a big valley outperformed state-of-the-art algorithms on well-known benchmark functions.
|
Research Products
(17 results)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
[Presentation] A New Framework taking account of Multi-funnel Functions for Real-coded Genetic Algorithms2011
Author(s)
Uemura, K., Kinoshita, S., Nagata, Y., Kobayashi, S. and Ono, I
Organizer
Proc. 2011 IEEE Congress on Evolutionary Computation (CEC), In CD-ROM, 8 pages
Place of Presentation
New Orleans
Year and Date
20110605-08
-
