María Botón-Fernández, Francisco Prieto-Castrillo, Miguel A. Vega-Rodríguez
Capítulo de libro
Applications of Evolutionary Computation. LNCS, Vol. 7835.
Springer-Verlag, Berlin Heidelberg, Germany, 2013, pp. 82-91.
Small-world property is found in a wide range of natural, biological, social or transport networks. The main idea of this phenomenon is that seemingly distant nodes actually have very short path lengths due to the presence of a small number of shortcut edges running between clusters of nodes. In the present work, we apply this principle for solving the resources selection problem in grid computing environments (distributed systems composed by heterogeneous and geographically dispersed resources).
The proposed model expects to find the most efficient resources for a particular grid application in a short number of steps. It also provides a self-adaptive ability for dealing with environmental changes. Finally, this selection model is tested in a real grid infrastructure. From the results obtained it is concluded that both a reduction in execution time and an increase in the successfully completed tasks rate are achieved.