Towards a coordination model for parallel cooperative P2P multi-objective optimization

Abstract

Existing Dispatcher-Worker Peer-to-Peer (P2P) computing environments are well-suited for multi-parameter applications. However, they are limited regarding the parallel computing where the generated tasks need to communicate. In this paper, we investigate that limitation and propose a coordination model for parallel P2P multi-objective optimization (MOO). The model has been implemented on top of the XtremWeb middleware. Then, it has been experimented on a combinatorial optimization application: a parallel branch-and-bound algorithm applied to the multi-objective (MO) Flow-Shop scheduling problem. The preliminary results obtained on a network of 120 heterogeneous PCs demonstrate the efficiency of the proposed approach. © Springer-Verlag Berlin Heidelberg 2005.