2022 Fiscal Year Final Research Report
Development of Multi-objective Evolutionary Computation Algorithms Based on Adaptive Operator Slection and Dynamic System Learning
| Project/Area Number |
20K11997
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 61040:Soft computing-related
|
|---|
| Research Institution | Tokyo University of Science |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2020-04-01 – 2023-03-31
|
| Keywords | 多目的進化計算 / 適応的交叉 / 航空交通 |
|---|
| Outline of Final Research Achievements |
Each step of the evolutionary computation can be modularised and interchangeable, and a performance improvement can be expected if a good combination is made according to the problem. However, since it is difficult to determine the optimal combination in advance, this study carried out (i) a reduction method of the hyperparameters of the adaptive approach, (ii) a dimensionality reduction method of the design parameters, and (iii) a parallel implementation of a real problem.
As a result, the minimum selection probability, one of the parameters of the adaptive crossover approach, was reduced. As a dimensionality reduction method, a genetic operator based on principal component analysis was proposed and its effectiveness was confirmed.
As a real-world problem, a multi-objective delay minimisation problem for air traffic flow over Japan was tackled and optimised using a parallel implementation.
|
| Free Research Field |
進化計算
|
| Academic Significance and Societal Importance of the Research Achievements |
進化計算は対象をブラックボックスとして取り扱えるため，産業界でも広く使われている．アルゴリズムの適用自体は比較的容易であるものの，問題の規模が増大するにつれて，更なる効率化が求められている．
本研究は進化計算アルゴリズムの更なる効率化を目指したものである．問題に応じて最適な遺伝オペレータを選択する手法や，主成分分析の基づく遺伝オペレータに焦点を当て，その有効性とパラメータの削減を実現した．また，実問題として航空交通流を考えその並列実装を行った．これらの成果は，大規模な最適化問題に取り組む際に適用できる基盤技術の一部となると考えられる．その適用範囲は広く，様々な分野での応用が期待できる．
|
