Artificially inducing environmental changes in evolutionary dynamic optimization
Biological and artificial evolution can be speeded up by environmental changes. From the evolutionary computation perspective, environmental changes during the optimization process generate dynamic optimization problems (DOPs). However, only DOPs caused by intrinsic changes have been investigated in the area of evolutionary dynamic optimization (EDO). This paper is devoted to investigate artificially induced DOPs. A framework to generate artificially induced DOPs from any pseudo-Boolean problem is proposed. We use this framework to induce six different types of changes in a 0–1 knapsack problem and test which one results in higher speed up. Two strategies based on immigrants, which are used in EDO, are adapted to the artificially induced DOPs investigated here. Some types of changes did not result in better performance, while some types led to higher speed up. The algorithm with memory based immigrants presented very good performance.
Citation : Tinos, R. and Yang, S. (2016) Artificially inducing environmental changes in evolutionary dynamic optimization. Proceedings of the 14th International Conference on Parallel Problems Solving from Nature (PPSN XIV), Lecture Notes in Computer Science, vol. 9921, pp. 225-236
ISBN : 9783319458229
ISSN : 0302-9743
Research Group : Centre for Computational Intelligence
Research Institute : Institute of Artificial Intelligence (IAI)
Peer Reviewed : Yes