Abstract
An alternative to the traditional random method of generating the initial population of genetic algorithms is introduced. The orthogonal arrays of the Taguchi method are used to create a population of chromosomes. The performance of these populations is compared empirically with populations created in the conventional random way. Just one test function is consider here which adequately illustrates the strengths and weaknesses of the new techniques The addition of Gaussian noise to the new initialization routine is shown to further improve the results.
Original language | English |
---|---|
Title of host publication | Proceedings 20th European Conference on Modelling and Simulation |
Editors | Wolfgang Borutzky, Alessandra Orsoni, Richard Zobel |
Publisher | ECMS |
Number of pages | 5 |
ISBN (Print) | 0-9553018-0-7 |
Publication status | Published - 2006 |
Event | 20th European Conference on Modelling and Simulation - Bonn, Germany Duration: 28 May 2006 → 31 May 2006 |
Conference
Conference | 20th European Conference on Modelling and Simulation |
---|---|
Abbreviated title | ECMS 2006 |
Country/Territory | Germany |
City | Bonn |
Period | 28/05/06 → 31/05/06 |