Name
Abstract | ||
|---|---|---|
This paper outlines the application of evolutionary search methods to problems in aeronautical design optimisation. The procedures described are based on the genetic algorithm (GA) and may be applied to other areas. Although easy to implement, a simple genetic algorithm is often found in applications to be of low efficiency and to suffer from premature convergence. To improve performance, two alternative strategies are investigated. In the first, a learning classifier scheme is used to tune the GA for a particular class of problems. The second strategy uses a parallel distributed genetic algorithm supervised by single or competing agents. The implementation of each procedure, and results for typical design problems are outlined. The agent supervised distributed genetic algorithm is found to provide a model with a very high degree of adaptibility, and to lead to considerably improved efficiency. |
Year | DOI | Venue |
|---|---|---|
2000 | 10.1007/3-540-45561-2_35 | EvoWorkshops |
Keywords | Field | DocType |
high degree,simple genetic algorithm,typical design problem,genetic algorithm,alternative strategy,evolutionary search method,improved efficiency,low efficiency,supervised evolutionary methods,aerodynamic design optimization,classifier scheme,aeronautical design optimisation,design optimization,premature convergence | Algorithm design,Premature convergence,Computer science,Meta-optimization,Genetic representation,Artificial intelligence,Cultural algorithm,Classifier (linguistics),Population-based incremental learning,Machine learning,Genetic algorithm | Conference |
Volume | ISSN | ISBN |
1803 | 0302-9743 | 3-540-67353-9 |
Citations | PageRank | References |
0 | 0.34 | 4 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| D. J. Doorly | 1 | 5 | 2.53 |
| S. Spooner | 2 | 0 | 0.34 |
| Joaquim Peiró | 3 | 39 | 7.28 |