Estimación y validación de distribuciones de longitud de ejecución para CHC y PBIL aplicados al problema de la selección de la solución deseada

View/Open
Document typeResearch report
Defense date2008-10
Rights accessOpen Access
All rights reserved. This work is protected by the corresponding intellectual and industrial
property rights. Without prejudice to any existing legal exemptions, reproduction, distribution, public
communication or transformation of this work are prohibited without permission of the copyright holder
Abstract
Devising a model to figure out metaheuristics performances when applied to high demanding problems is paramount. Such a model would allow to foresee the computational time required for a given output quality. In previous works we showed that the performance of CHC and PBIL evolutive algorithms fit in a Gamma distribution when applied to solve the Root Identification Problem for search domains bounded by 2$20$ instances. Here we conduct experimental studies to verify that the Gamma model applies to search spaces of up to 2$^50$. Then we apply the model to estimate the runtime for search spaces of size up to 2$^100$. Finally we run the algorithms and validate the model by comparing the actual runtime values to those yielded by the model.
CitationYeguas, E., Joan-Arinyo, R. "Estimación y validación de distribuciones de longitud de ejecución para CHC y PBIL aplicados al problema de la selección de la solución deseada". 2008.
Is part ofLSI-08-27-R
Collections
Files | Description | Size | Format | View |
---|---|---|---|---|
R08-27.pdf | 1,095Mb | View/Open |