This work is being carried out in the department of computing science and mathematics at the University of Stirling, Scotland. It introduces novel crossover techniques for genetic algorithms (GAs). GAs are a stochastic search mechanism inspired by Darwin's theory of evolution, and in our work we are applying these techniques to dynamic systems in a time series environment. At the moment the application area under investigation is scheduling the nematode Steinernema feltiae as a bio-control mechanism for protecting mushroom crops from sciarid flies.
The results derived to date in this area seems to highlight the effectiveness of the approaches created by this work in deriving optimal bio-control strategies. Our results show an improvement of our approaches over Uniform Crossover (UC), which is a industry standard approach to crossover.
Although the scheduling of nematodes as a bio-control strategy has been the area of investigation to date, we are about to test our approaches in the area of optimal scheduling for cancer treatments in a chemotherapy model. Analysis of the results produced in these tests will hopefully show the effectiveness of the approaches constructed in this project.
Main Page | About the project | Related work | Presentations of work | Publications