Abstract
This article presents several results obtained in the field of CFD optimization using Genetic Algorithms (GAs). We first introduce a new approach based on classical GAs, namely Hierarchical Genetic Algorithms (HGAs). HGAs are explained in details, and the advantages of a multi-layered hierarchical topology are clearly shown. The hierarchical topology is an excellent compromise to the exploration/exploitation dilemma often encountered by GAs. Another innovation is the introduction of multiple models for optimization problems, within the frame of an HGA. We show that with such an architecture, it is possible to use a mix of simple models (e.g. coarse meshes) that are very fast and more complex models (with slower solvers), and still achieve the same quality as that obtained with only complex models. The different concepts presented in this paper are then illustrated via experiments on a Computational Fluid Dynamics problem, namely a nozzle reconstruction. The Mach number distribution along the nozzle is solved by a time marching technique using a CUSP scheme with an iterative solver. We show the results obtained and we compare the CPU time requirements for traditional GAs, Hierarchical GAs with single model and Hierarchical GAs with multiple models. And most importantly, we show how Hierarchical GAs deal with approximate models when searching for a global solution.
Original language | English |
---|---|
Title of host publication | European Congress on Computational Methods in Applied Sciences and Engineering, ECCOMAS 2000 |
Place of Publication | Barcelona, Spain |
Publisher | EccOMAS |
Pages | 1-18 |
Number of pages | 18 |
ISBN (Print) | 8489925704, 9788489925700 |
Publication status | Published - 2000 |
Event | European Congress on Computational Methods in Applied Sciences and Engineering, ECCOMAS 2000 - Barcelona, Spain Duration: 11 Sept 2000 → 14 Sept 2000 |
Other
Other | European Congress on Computational Methods in Applied Sciences and Engineering, ECCOMAS 2000 |
---|---|
Country/Territory | Spain |
City | Barcelona |
Period | 11/09/00 → 14/09/00 |
Keywords
- Computational fluid dynamics
- Genetic algorithms
- Hierarchical models
- Optimization
- Parallelism