A study on large-scale optimization by efficient search methods using approximation models with low accuracy
Project/Area Number |
20500138
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Single-year Grants |
Section | 一般 |
Research Field |
Intelligent informatics
|
Research Institution | Hiroshima City University |
Principal Investigator |
|
Project Period (FY) |
2008 – 2011
|
Project Status |
Completed (Fiscal Year 2011)
|
Budget Amount *help |
¥3,770,000 (Direct Cost: ¥2,900,000、Indirect Cost: ¥870,000)
Fiscal Year 2011: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2010: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2009: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Fiscal Year 2008: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
|
Keywords | 最適化アルゴリズム / 進化的計算 / Differential Evolution / 低精度近似モデル / 近似モデル / 機械学習 / 制約付き最適化 / Expensive Optimization / 低精度折似モデル / Expensive optimization |
Research Abstract |
In this study, we proposed and improved "estimated comparison method", which reduces the number of function evaluations by omitting the function eval-uations when the result of comparison can be judged by approximation values, in order to realize a general purpose efficient optimization algorithm. In this study, it is shown that the efficient general purpose algorithm can be built by introducing the potential model that is an approximation model with low accuracy, introducing the margin error parameter and the congestion parameter, proposing a new way of parameter settings for differential evo-lution and combining the method with the epsilon constrained method.
|
Report
(6 results)
Research Products
(96 results)