Omni-directional RND Optimisation using Differential Evolution: In-depth Analysis via High Throughput Computing



The Radio Network Design (RND) constitutes an important class of problems, particularly in the planning of wireless communication networks. RND problems are challenging to tackle since they fall in the NP-hard class of optimisation problems. In this paper, we assess the viability of adapting the Differential Evolution (DE) algorithm to a wide-scale real world RND problem. To fulfil the high computational demands of the DE approach, we resort to a pool of more than 150 non-dedicated machines, whose CPU cycles are scavenged through a high throughput system. Our results show that DE is a viable approach for RND problems if proper computing power is available.


Differential evolution, high throughput computing, Radio Network Design


Differential Evolution


EPIA 2007 - Portuguese Conference on Artificial Intelligence., December 2007

PDF File

Cited by

No citations found