Efficient parallel LAN/WAN algorithms for optimization. The mallba project

Abstract

The mallba project tackles the resolution of combinatorial optimization problems using algorithmic skeletons implemented in C ++ . mallba offers three families of generic optimization techniques: exact, heuristic and hybrid. Moreover, for each technique, mallba provides three different implementations: sequential, parallel for local area networks, and parallel for wide area networks. This paper explains the architecture of the mallba library, presents some of the implemented skeletons, and offers several computational results to show the viability of the approach. In our conclusions we claim that the design used to develop the optimization techniques is general and efficient at the same time, and also that the resulting skeletons can outperforms existing algorithms on a plethora of problems.

Dynamic Load Balancing on Dedicated Heterogeneous Systems

“…The library has been implemented in a way that does not require changing any line of code in existing programs, thus minimizing code intrusion. All that is required is to use three new functions: We validated our proposal on three test problems: matrix product [5], the Jacobi method for solving linear systems [5] and resource allocation optimization via dynamic programming algorithms [6]. The computational results show that the benefits yielded by using our balancing library offer substantial time reductions in every case.…”

Meta-heuristics for Grid Scheduling Problems

“…For instance, Mallba library [1], Paradiseo [14] and EasyLocal++ [45] are such libraries. These libraries can be easily used for Grid scheduling problem; for instance, the meta-heuristic approaches in Xhafa et al [16,72,73] use skeletons defined in Mallba Library.…”

“…This method has been proposed for Grid scheduling problem by Abraham et al [1] and Yarkhan and Dongarra [75].…”

“…However, its mechanisms of tabu lists, aspiration criteria, intensification and diversification make it very powerful search algorithm. Abraham et al [1] also considered Tabu Search as candidate solution method for the problem. Ritchie [57] implemented the TS for the problem under ETC model and used it in combination with ACO approach.…”

“…GAs for Grid scheduling problems have been addressed by Abraham et al [1], Braun et al [9], Zomaya and Teh [81], Martino and Mililotti [46], Page and Naughton [53], Carretero and Xhafa [16], Gao et al [29], Xhafa et al [70,73].…”