On the Design of a Master-Worker Adaptive Algorithm Selection Framework - Archive ouverte HAL Access content directly
Conference Papers Year : 2018

On the Design of a Master-Worker Adaptive Algorithm Selection Framework

Abstract

We investigate the design of a master-worker schemes for adaptive algorithm selection with the following two-fold goal: (i) choose accurately from a given portfolio a set of operators to be executed in parallel, and consequently (ii) take full advantage of the compute power offered by the underlying distributed environment. In fact, it is still an open issue to design online distributed strategies that are able to optimally assign operators to parallel compute resources when distributively solving a given optimization problem. In our proposed framework, we adopt a reward-based perspective and investigate at what extent the average or maximum rewards collected at the master from the workers are appropriate. Moreover, we investigate the design of both homogeneous and heterogeneous scheme. Our comprehensive experimental study, conducted through a simulation-based methodology and using a recently proposed benchmark family for adaptive algorithm selection, reveals the accuracy of the proposed framework while providing new insights on the performance of distributed adaptive optimization algorithms.
Fichier principal
Vignette du fichier
dams-ea17.pdf (740.29 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-01643354 , version 1 (13-09-2021)

Identifiers

Cite

Christopher Jankee, Sébastien Verel, Bilel Derbel, Cyril Fonlupt. On the Design of a Master-Worker Adaptive Algorithm Selection Framework. EA 2017 - 13th International Conference on Artificial Evolution, Oct 2017, Paris, France. pp.1-15, ⟨10.1007/978-3-319-78133-4_1⟩. ⟨hal-01643354⟩
279 View
23 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More