Scheduling is a very important problem in many real-world scenarios. In the case of supercomputers it is even more important because available resources are limited and expensive. The optimal use of supercomputer facilities is a critical question. We have found that the definitions of traditional scheduling problems do not provide an appropriate description for Supercomputer Scheduling (SCS). Thus, a new definition for this kind of problems is proposed. The research already done in the field of other scheduling problems can be modified to be applied in this new scenario. Nevertheless, new techniques can also be developed. Thus, we have proposed a theoretical framework to combine multi evolutionary algorithms called Multiple Offspring Sampling (MOS). We have used this approach to combine multiple codings and genetic operators in this scheduling problem. To summarise: first, we introduce a formal definition of supercomputer scheduling; second, we propose Multiple Offspring Sampling formalism; and third, we have carried out an experimental test to compare the performance of this formalism to solve SCS problems against traditional (non-combinatorial) techniques and single genetic algorithms. © 2008 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Latorre, A., Peña, J. M., Robles, V., & de Miguel, P. (2008). Supercomputer scheduling with combined evolutionary techniques. Studies in Computational Intelligence, 146, 95–120. https://doi.org/10.1007/978-3-540-69277-5_4
Mendeley helps you to discover research relevant for your work.