| Presentation | 2006-06-16 Fast Method of Genetic Algorithm Searching and its Application to Neural Network Training Moriyoshi MAEHSIRO, Hiroshi KINJO, Kunihiko NAKAZONO, Tetsuhiko YAMAMOTO, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Genetic algorithm (GA) is known to be one of the most powerful solution searching mechanism for nonlinear and multi-variable optimization problems. Generally, GA takes many long times to find the solutions and sometimes it cannot find the optimum solutions. In order to improve the searching performance, we propose a fast algorithm of GA and a mutation method. The fast algorithm is usage of a momentum offspring (MOS). The MOS is a individual not the crossover but by the best individuals between current and past generation. The MOS is considered it has higher probability for desired solution and the effect of MOS is fast searching of the optimum solution. Furthermore we proposed a constant range mutation (CRM) for the GA. The CRM is considered it has an effect of avoiding the ineffective individual production. We apply the GA with MOS and CRM to neural network training. Simulation shows proposed method has good training performances. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Genetic algorithm / Fast algorithm / Momentum offspring / Constant range mutation / Neural network training |
| Paper # | NC2006-36 |
| Date of Issue |
| Conference Information | |
| Committee | NC |
|---|---|
| Conference Date | 2006/6/9(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Neurocomputing (NC) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Fast Method of Genetic Algorithm Searching and its Application to Neural Network Training |
| Sub Title (in English) | |
| Keyword(1) | Genetic algorithm |
| Keyword(2) | Fast algorithm |
| Keyword(3) | Momentum offspring |
| Keyword(4) | Constant range mutation |
| Keyword(5) | Neural network training |
| 1st Author's Name | Moriyoshi MAEHSIRO |
| 1st Author's Affiliation | Graduate School of Mechanical Engineering, University of the Ryukyus() |
| 2nd Author's Name | Hiroshi KINJO |
| 2nd Author's Affiliation | University of the Ryukyus |
| 3rd Author's Name | Kunihiko NAKAZONO |
| 3rd Author's Affiliation | University of the Ryukyus |
| 4th Author's Name | Tetsuhiko YAMAMOTO |
| 4th Author's Affiliation | University of the Ryukyus |
| Date | 2006-06-16 |
| Paper # | NC2006-36 |
| Volume (vol) | vol.106 |
| Number (no) | 102 |
| Page | pp.pp.- |
| #Pages | 4 |
| Date of Issue | |