MALLBA: towards a combinatorial optimization library for geographically distributed systems
Enrique Alba, Filipe Guedes de Oliveira Almeida, Maria Josep Blesa, Carlos Cotta, Miguel Ángel Prieto Díaz, I. Dorta, Joaquim Gabarró, Jesús A. González, Coromoto León, Luz Marina Moreno de Antonio, Jordi Petit, José L. Roda, A. Rojas, Fatos Xhafa · LA Referencia (Red Federada de Repositorios Institucionales de Publicaciones Científicas) · 2001
Abstract—Problems arising in different areas such as numerical methods, simulation or optimization can be efficiently solved by parallel super-computing. How-ever, it is not always possible to buy and maintain par-allel super-computers. A geographically distributed network of PC clusters is an interesting low-cost al-ternative. The possibility of connecting different clus-ters of PCs through Internet opens a new approach to distributed and massive computing. The MALLBA project tackles the resolution of combinatorial opti-mization problems using algorithmic skeletons imple-mented in C++ under this approach. MALLBA offers three families of generic resolution methods: exact, heuristic and hybrid. Moreover, for each resolution method it offers three implementations: sequential, LAN and WAN. This paper surveys the current state of the MALLBA project.