|
| Engineering structural optimization using a cooperative optimization method based on evolution strategy and gaussian process |
| Received:April 08, 2012 Revised:June 28, 2012 |
| View Full Text View/Add Comment Download reader |
| DOI：10.7511/jslx201305003 |
| KeyWord:structural optimization design evolution strategy gaussian process approximation model |
| Author | Institution |
| 苏国韶 |
广西大学 土木建筑工程学院 工程防灾与结构安全教育部重点实验室, 南宁 |
| 武振兴 |
广西大学 土木建筑工程学院 工程防灾与结构安全教育部重点实验室, 南宁 ;中建钢构有限公司, 深圳 |
| 燕柳斌 |
广西大学 土木建筑工程学院 工程防灾与结构安全教育部重点实验室, 南宁 |
|
| Hits: 3417 |
| Download times: 2342 |
| Abstract: |
| The optimization of complex structures with large number of degrees of freedom using bionic optimization will have the huge computational volume and will be very time consuming.For reducing the computational burden of bionic optimization,a novel cooperative optimization method (CMAES-GP-FEM) based on Covariance Matrix Adaptation Evolution Strategy (CMAES) algorithm,Gaussian process (GP) machine learning and Finite Element Method (FEM) is proposed.In the proposed method,CMAES algorithm is applied to search the globe optimization solution.GP model is applied to approximate the fitness function when the CMAES algorithm going into the statue of searching local minimum.Thus,the computational cost induced by reducing the numbers of finite element analysis during the local searching process.Compared with the traditional optimization and bionic optimization method for structural optimization,the proposed method is much more economical to achieve the global solution. |