2015 Fiscal Year Final Research Report
The development of the differential evolution programming using Inter-symbol distance
Project/Area Number |
26730131
|
Research Category |
Grant-in-Aid for Young Scientists (B)
|
Allocation Type | Multi-year Fund |
Research Field |
Soft computing
|
Research Institution | Hiroshima City University |
Principal Investigator |
|
Project Period (FY) |
2014-04-01 – 2016-03-31
|
Keywords | Differential evolution / Genetic Programming / 進化的計算 / 最適化 |
Outline of Final Research Achievements |
In this study, we have developed a novel evolutionary algorithm for tree discovery based on the Differential Evolution (DE). In order to apply DE’s genetic manipulation to tree structure, we imparted neighborhood structure to the symbols contained in trees. The distance between symbols are defined using the neighborhood structure and differential operation can be performed in the discrete symbol space. For generation model to evolve both tree structure and neighborhood structure of symbols, we have proposed two models: the self adaptive model and the coevolutionary model. To evaluate the performance of the proposed methods, we conducted experiments using benchmark problems. Through experimental results, we confirmed that the proposed method outperforms genetic programming in symbolic regression problem and even-parity problem.
|
Free Research Field |
進化的計算
|