2021 Fiscal Year Final Research Report
Evolutionary Constrained Multiobjective Optimization Utilizing Constraint Satisfaction Values of Solutions
| Project/Area Number |
19K20346
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 61030:Intelligent informatics-related
|
|---|
| Research Institution | Shinshu University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2019-04-01 – 2022-03-31
|
| Keywords | 進化計算 / 制約付き多目的最適化 / 制約許容量 |
|---|
| Outline of Final Research Achievements |
To improve the search performance of evolutionary computation for solving constrained multi-objective optimization problems, it is important to elicit more information from candidate solutions and utilize it during the solutions search. This study newly introduced “constraint satisfaction values” evaluating how a solution is far from the boundary of a constraint when the solution satisfies the constraint. To utilize the values, this study focused on problems that the pareto optimal solutions are on a boundary of the constraints, and a mating method selecting a pair of parent solutions balancing the constraint violation value and constraint satisfaction value of solutions on the constraints is proposed and verified the effectiveness.
|
| Free Research Field |
進化計算
|
| Academic Significance and Societal Importance of the Research Achievements |
これまでの進化計算では，解探索中に生成された制約違反解は最終的に解として認められないため，制約充足解よりも解探索に用いられにくい．しかし，現実世界の多目的最適化問題では制約条件が多数設定されることが多く，また，制約境界に近づくほど目的が最適化される傾向があるため，解探索中には多数の制約違反解が生成される．そのため，制約違反解の活用による解探索の効率化は実応用に向けて重要な検討である．また，本研究で提案する制約許容量は制約違反解の活用手段をさらに増やすことが期待される．
|
