Title | A ``Futurist'' approach to dynamic environments |
Publication Type | Conference Paper |
Year of Publication | 2001 |
Authors | van Hemert, JI, Van Hoyweghen, C, Lukschandl, E, Verbeeck, K |
Conference Name | Proceedings of the Workshops at the Genetic and Evolutionary Computation Conference, Dynamic Optimization Problems |
Publisher | Morgan Kaufmann Publishers, San Francisco |
Editor | Branke, J, B{\"a}ck, T |
Keywords | dynamic problems |
Abstract | The optimization of dynamic environments has proved to be a difficult area for Evolutionary Algorithms. As standard haploid populations find it difficult to track a moving target, diffKerent schemes have been suggested to improve the situation. We study a novel approach by making use of a meta learner which tries to predict the next state of the environment, i.e. the next value of the goal the individuals have to achieve, by making use of the accumulated knowledge from past performance. |
Full Text |